![[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}
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,
Cell and Molecular Microbiology Division, Institute of Microbiology, Academy of Sciences of the Czech Republic, Víde
ská 1083, 142 20 Prague 4, Czech Republic
Received 17 October 2006/ Accepted 19 March 2007
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
stkP mutant strains by microarray technology. This analysis revealed that StkP positively controls the transcription of a set of genes encoding functions involved in cell wall metabolism, pyrimidine biosynthesis, DNA repair, iron uptake, and oxidative stress response. Despite the reduced transformability of the stkP mutant, we found that the competence regulon was derepressed in the stkP mutant under conditions that normally repress natural competence development. Furthermore, the competence regulon was expressed independently of exogenous competence-stimulating peptide. In summary, our studies show that a eukaryotic-type serine/threonine protein kinase functions as a global regulator of gene expression in S. pneumoniae. |
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Prokaryotic STPKs have been shown to regulate various cellular functions, such as developmental processes (47, 37, 49), secondary metabolism (77), stress responses (27, 50), metabolic processes (48, 66), biofilm formation (27), and also virulence in streptococci (18, 28, 64), Yersinia pseudotuberculosis (21), Pseudomonas aeruginosa (80), and Mycobacterium tuberculosis (14, 79). Although these studies established the function of protein kinases, only several target substrates have been identified to date (16, 31, 42, 48, 55, 64, 66, 78). The result of posttranslational modification of a substrate on its function has been demonstrated in just a minority of cases (49, 66).
In contrast to the case with many microorganisms, Streptococcus pneumoniae and other streptococci contain a single gene encoding an STPK, which is genetically linked to a protein phosphatase-encoding gene (18, 28, 64). As a result, this feature makes them a more straightforward and practical model for defining the regulatory functions provided by these proteins.
We recently characterized the biochemical properties of both the protein kinase StkP and the protein phosphatase PhpP of S. pneumoniae and showed that the autophosphorylated form of StkP is a substrate for PhpP. These results suggest that StkP and PhpP could operate as a functional pair in vivo (55). It has been demonstrated that such functional coupling could play a role in the control of kinase activity (4). An analysis of phosphoproteome maps of both wild-type and stkP-null mutant strains revealed that one of the endogenous substrates of StkP is phosphoglucosamine mutase GlmM (55). GlmM is an essential enzyme involved in the biosynthesis of UDP-N-acetylglucosamine, a key common precursor to cell envelope components (41). The phosphorylation of GlmM is necessary for its activity in Escherichia coli and is achieved during an autophosphorylation process (29). On the other hand, the phosphorylation of GlmM by StkP is not essential for its enzymatic activity since the S. pneumoniae stkP mutant is viable. In addition to GlmM, we have identified the 
-subunit of RNA polymerase as another substrate of StkP in in vivo-labeling experiments, although the significance of its modification is unclear (55).
It has also been shown that the mutation of stkP attenuates virulence in both pneumonia and bacteremia models of infection, indicating that StkP plays an important role for bacterial survival in vivo (18). A recent report by Oggioni et al. provided evidence that the expression of stkP was significantly increased in vivo in both the blood and tissues of infected mice (56). During the infection process, S. pneumoniae is exposed to various stress factors, including nutritional deprivation, temperature shift, pH changes, increased osmolality, and reactive oxygen species generated by host phagocytes. Due to the transmembrane topology of StkP and the presence of an extracellular sensor domain containing reiterated PASTA (penicillin binding protein and Ser/Thr protein kinase-associated domain) signature sequences (81), we hypothesized that StkP could transmit environmental cues into the cell.
Here we report the growth characteristics of an S. pneumoniae stkP mutant under different stress conditions. These results suggest that the expression of StkP can contribute to the resistance of S. pneumoniae to environmental stresses. Microarray analysis was used to compare the expression profile of the stkP mutant with that of the wild type. This analysis revealed that the stkP mutation is broadly pleiotropic and affects the transcription of several sets of important genes. Several functional gene categories have been identified that could account for a reduced stress response and attenuated virulence. Furthermore, we show that StkP of S. pneumoniae plays a role in the maintenance of low expression levels of competence genes under conditions that do not support competence development. Finally, we propose a hypothesis to describe the putative mechanism of competence deficiency in an stkP mutant.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
stkP (like CP1015 but stkP::Cm [55]), were grown at 37°C in casein tryptone (CAT) medium supplemented with 0.2% glucose and 1/30 volume 0.5 M K2HPO4, pH 7.5. DNA from strain CP1016 (rif-23) was used as donor DNA for transformation assays (45). The following antibiotics were added when necessary at the indicated concentrations: chloramphenicol, 10 µg ml–1; and rifampin, 1 µg ml–1. Construction of S. pneumoniae stkP mutant. The deletion of the stkP gene was achieved by transforming an S. pneumoniae wild-type strain with a vectorless DNA fragment consisting of stkP downstream and upstream regions of homology and a cat cassette replacing the stkP coding region as previously described (55). Viable chloramphenicol-resistant clones arising from a double-crossover event were readily obtained after S. pneumoniae transformation. The resulting clones were further examined for successful allelic exchange by diagnostic PCR and Southern hybridization. The junctions between exogenous and chromosomal DNA were verified by sequencing.
In vitro stress experiments.
To compare the effect of temperature stress on the wild-type and stkP mutant strains, we used the same number of cells (4 x 104 CFU ml–1) to inoculate CAT medium prewarmed to 37 and 40°C. Cultures were incubated statically under aerobic conditions at the indicated temperatures. At 1-h intervals, samples were removed and their optical densities (OD) were determined at a wavelength of 400 nm.
The sensitivity of cells to H2O2 was tested by exposing aliquots of cultures (107 CFU ml–1; OD400 = 0.4) grown in CAT medium at 37°C to 40 mM H2O2 for 5, 10, 15, and 20 min at 37°C. Viable cells were counted by plating them onto agar plates before and after exposure to H2O2, and results were expressed as percentages of survival.
Adaptability of the wild-type and isogenic CP1015stkP mutant to osmotic stress was evaluated by monitoring bacterial growth in CAT medium containing 400 mM NaCl. To initiate the growth experiments, bacteria grown to early exponential phase (OD400 = 0.2) were inoculated into prewarmed CAT medium with or without NaCl. At 1-h intervals, bacterial growth was monitored by measuring the OD400.
To study the role of StkP in acid tolerance, we used equal numbers of cells (4 x 104 CFU ml–1) of both the wild-type strain and the stkP mutant to inoculate CAT medium adjusted to pHs 6.5 and 7.5. At 1-h intervals, samples were withdrawn and the OD400 values were determined.
RNA extraction. For an analysis of the transcriptome and quantitative real-time PCR, RNA was isolated from 50 ml of pneumococcal culture grown at 37°C in CAT medium to an OD400 of 0.4. Total bacterial RNA was extracted with hot acid phenol according to the method of Cheng et al. (10). After preliminary purification and precipitation with isopropanol, the crude RNA was treated with DNase I (Ambion) and purified further using RNeasy columns (QIAGEN) according to the manufacturer's protocol. The typical yield from 50 ml of cells was approximately 200 µg of purified RNA.
Reverse transcription and microarray hybridization. Commercial S. pneumoniae whole-genome microarray slides from Eurogentec (Belgium) were used in this study. The full genome array consists of amplicons representing segments of 2,085 open reading frames from S. pneumoniae reference strain TIGR4 spotted in duplicate on a glass slide.
Ten micrograms of purified RNA samples was reverse transcribed and differentially labeled with Cy3-dCTP and Cy5-dCTP (Amersham Biosciences) in the presence of random hexamer primers (Amersham Biosciences) and SuperScript II RNase H– RT (Invitrogen) in a total volume of 20 µl. The reaction mixture was incubated at 42°C for 1.5 h. Dye swap labeling was performed systematically for all labeled RNA samples. Pairs of Cy3- and Cy5-labeled cDNA samples were mixed and hybridized to the array at 37°C for 15 h. After hybridization, the slides were washed in 0.2x SSC (1x SSC is 0.15 M NaCl plus 0.015 M sodium citrate)-0.1% sodium dodecyl sulfate for 5 min, followed by a wash in 0.2x SSC for 5 min.
Data analysis.
The arrays were scanned using the GeneTAC UC-4 scanner (Genomic Solutions). TIGR Spotfinder 2.2.3 (71) was used for determining the average signal intensity and the local background for each spot. The Cy3/Cy5 fluorescence ratios were log2 transformed and normalized by LOWESS normalization in the TIGR microarray data analysis system (MIDAS) 2.17 (71). In total, six hybridizations using samples from three independent bacterial cultures and RNA extractions were performed and the average was calculated from the independent experiments. The statistical significance of the expression ratios was determined by applying the Student t test. Genes whose expression ratios changed by a factor of at least 2 and had P values of <0.01 were considered significantly regulated by StkP (95 genes). In order to verify the significance of regulation by a different statistical method, the data matrix of log2-transformed gene expression ratios of all replicates was imported into TIGR MultiExperiment Viewer (MeV) 4.0 (71) and analyzed by a one-class significance analysis of microarrays (SAM) test (76). SAM statistics were based on all 4,096 possible permutations, and the threshold delta value ( = 0.91) was based on delta versus the percent false discovery rate table reported by SAM to avoid any falsely significant genes. SAM identified 208 significantly regulated genes and confirmed all 97 genes identified by Student t test and expression ratio threshold.
Validation of array data. The significance of signals obtained from microarray hybridizations was verified by quantitative real-time PCR (qRT-PCR) (see below) using Sybr green I JumpStart Taq ReadyMix kit (Sigma). Six derepressed genes (spr0013, spr0857, spr1109, spr1724, spr1864, and spr2041), eight repressed genes (spr0096, spr0867, spr0934, spr1057, spr1154, spr1320, spr1684, and spr2021), and two stably expressed genes (spr0708 and spr1757) were chosen for this confirmatory study. qRT-PCRs were performed with the same RNA samples that were used for microarray analysis. For each gene, three biological replicates with three technical replicates each were used.
Competence gene expression. Cells were grown in CAT medium at 37°C until an OD400 of 0.4 before exposure to competence-stimulating peptide (CSP) (250 ng ml–1). At specified intervals (8 and 13 min) after the addition of CSP, total RNAs were extracted (see above). Sample "0" was taken before the addition of CSP. Reverse transcription was performed as described above. For the expression profile analysis of competence genes, qRT-PCR was performed (see below) on a select group of competence genes: spr0013, spr0020, spr0043, spr0857, spr1724, spr1864, spr2041, and spr2043.
qRT-PCR.
qRT-PCRs were carried out in a Mxp3000 LightCycler instrument (Stratagene). All primers were calculated using Primer3 services at http://frodo.wi.mit.edu. PCR products were approximately 100 bp in length. A master mix of the following reaction components was prepared to the following end concentrations: 6.9 µl water, 1 µl forward primer (0.5 µM), 1 µl reverse primer (0.5 µM), 0.2 µl internal Rox reference dye, and 9.9 µl Sybr green JumpStart Taq ReadyMix (Sigma). One microliter of cDNA (50 ng reverse-transcribed total RNA) was added as a PCR template. Amplification was performed using the following cycling protocol: 3 min of initial denaturation at 95°C, followed by a three-step profile consisting of 30 s of denaturation at 95°C, 25 s of annealing at 60°C, and 25 s of extension at 72°C for a total of 40 cycles. Each qRT-PCR was performed in triplicate using RNA from two independent isolations, and no-reverse-transcriptase controls were also included. The change in expression of each gene in the wild-type and CP1015stkP mutant strain was calculated according to the equation of Pfaffl (62). Values were normalized to the reference gyrA gene (spr1099) encoding the DNA gyrase subunit A. Melting curve analysis was also performed to evaluate PCR specificity and resulted in single primer-specific melting temperatures. For a list of the primers used, see Table S1 in the supplemental material.
Complementation experiment. An S. pneumoniae strain with an ectopic copy of the pcsB gene was constructed by transformation with a synthetic linear PCR amplicon kindly provided by M. Winkler (51). A bgaA::[kan-t1t2-PfcsK]-pcsB+ fragment was inserted into the dispensable bgaA chromosomal gene. The presence of the desired mutation was verified by screening for changes in the sizes of fragments amplified from chromosomal DNA using the flanking primers.
Competence assay. The transformation was performed as described previously (35). Briefly, cells were grown in CAT medium at 37°C to OD400 values of 0.1 and 0.4. Appropriate aliquots were withdrawn, incubated in the presence or absence of CSP (250 ng ml–1), and mixed with rif-23 donor DNA (1 µg ml–1). Further incubation was carried out at 37°C for 1 h before plating in CAT agar containing 1 µg Rif ml–1. Serial dilutions of transformed cultures were plated, and transformation efficiencies were calculated as the ratio of the viable counts on plates with and without rifampin. Results are expressed as the mean counts from two independent experiments.
|
RESULTS AND DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
In vitro growth phenotype of the stkP mutant.
Prior to studying the effect of the stkP mutation on the global expression profile of S. pneumoniae, we first examined the growth characteristics of the stkP-null mutant in vitro. The stkP mutant strain displayed small but significant changes in growth properties under nonstress conditions. A comparison of the growth levels of the mutant and parent strains in CAT medium at 37°C showed that the mutant had a significantly reduced growth rate compared to that of the parent strain in the early exponential phase (Fig. 1A). The calculated doubling time of the mutant was increased to 55 min compared to 35 min for the parent strain. In addition, cultures of the mutant strain exhibited an increased rate of autolysis at the stationary phase of growth and did not achieve the same final optical density as the parent strain. The extended lag phase and slower doubling time in the early exponential phase and the recovery of growth with a rate approaching that of the wild type in the mid-exponential phase, which we observed in our studies, are reminiscent of those observed for protein kinase mutants in other streptococci (27, 28, 64).
|
stkP mutant at an elevated temperature. In contrast to the parent strain, which was unaffected at 40°C, the growth of the mutant strain was almost completely impaired (Fig. 1A). This result suggests that StkP regulates genes involved in adaptation to high temperatures.
Sensitivity of the stkP mutant to oxidative stress.
To determine whether the stkP strain has an altered sensitivity to oxidative stress, its ability to survive H2O2- and superoxide (O2–)-induced oxidative stress was examined and compared with that of the parental strain. Decreased survival of the stkP mutant, compared to that of the wild type, was observed at 40 mM H2O2 (Fig. 1B). After 15 min of treatment with 40 mM hydrogen peroxide, 96% of the stkP cells were killed, whereas 69% of the wild-type cells survived. Both the wild-type and mutant strains were also exposed to 50 mM paraquat over a 30-min period. In contrast to the results of experiments using hydrogen peroxide as an oxidative agent, both the wild type and the stkP mutant exhibited virtually the same sensitivity to superoxide (data not shown). This sensitivity to oxidative killing might be a major contributor to the attenuated virulence of the stkP mutant (18), since such cells would be less likely to survive attacks by the innate host defense system.
The stkP mutant strain is more susceptible to osmotic stress.
To further characterize the ability of the stkP mutant strain to cope with stress conditions, its growth was measured under high osmolality. Growth of the wild-type strain in CAT medium containing 400 mM NaCl was impaired (Fig. 1C). The stkP mutant strain exhibited no discernible growth, and virtually all of the cells lysed at the end of cultivation. Hence, the stkP mutant has a reduced tolerance to osmotic stress.
The stkP mutant strain is less tolerant to acidic pH.
A comparison of the growth of the wild-type and mutant strains under mild acidic conditions (pH 6.5) showed that the stkP mutant is much less tolerant to an acidic pH (Fig. 1D). In contrast to that of the parent strain, which was only slightly affected, the growth of the mutant strain was significantly slower with an extended lag phase. A similar effect of Ser/Thr protein kinase deletion on acid tolerance has been observed in Streptococcus mutans (27). Taken together, these data indicate that the stkP mutation confers sensitivity to a variety of stressors, suggesting that it very likely affects numerous different cellular functions.
Global transcription profile of the stkP mutant.
To begin to address the genetic basis of these phenotypes, we performed an analysis of global transcription of the stkP mutant compared to the wild-type parent in the exponential phase (OD400 = 0.4) in CAT. Custom S. pneumoniae whole-genome microarray slides from Eurogentec (Materials and Methods) were used throughout this study. In total, six independent hybridizations using RNA samples from three separate bacterial cultures were performed for both strains. After background subtraction and normalization, the data were processed and statistically evaluated (Materials and Methods). Among 2,085 genes evaluated, changes in relative transcript amounts of at least twofold were observed for 95 genes. Under the conditions used, more than 4% of S. pneumoniae genes were identified as transcriptionally affected, indicating a global regulatory function for the StkP signaling protein. Complete lists of genes with affected transcript levels are given in Tables 1 and 2. The majority of altered genes could be clustered into several different functional categories.
|
|
stkP mutant (decrease of 22-fold) compared to the presence of the parent strain (Table 1). stkP is located downstream of, and overlaps with, phpP, a gene coding for StkP cognate phosphatase (55). The presence of a transcriptional terminator following stkP suggested that stkP is the last gene in the operon (18). Thus, the polar effect resulting from stkP replacement is unlikely. Furthermore, the level of transcription from the phpP (spr1578) and orf-1576 genes, which flank stkP, remained unaffected in the stkP mutant and its stkP+ parent, indicating a lack of polar effects on the surrounding genes. Validation of gene expression by qRT-PCR analysis. The RT-PCR approach was chosen as an independent method to validate the microarray results. We quantified transcripts of selected genes by using the same RNA samples that were used for the microarray assays. Average data from at least two experimental replicates are shown in Table 3. In each instance, the qRT-PCR results correlated well (R2 = 0.9) with those obtained from the microarrays (Fig. 2). qRT-PCR assays provided a higher dynamic range for relative gene expression than did the DNA microarrays, as indicated for spr1057 as well as spr0857 (celB), spr1684 (fatD), spr1724 (ssbB), and spr1864 (cglA). Both microarray and qRT-PCR analysis showed that the expression levels of recA and ciaH are unaltered in the stkP mutant.
|
|
stkP mutant.
Apart from a few exceptions, genes encoding proteins potentially involved in the regulation of cell wall biosynthesis were most strongly repressed (Table 1). spr0096 (11-fold decrease) encodes a hypothetical protein containing a LysM domain, one of the most common protein modules in bacterial cell surface proteins. LysM occurs most often in cell wall-degrading enzymes, where it is believed to act to anchor the catalytic domains to their substrates (30). Several proteins that contain the domain, including staphylococcal immunoglobulin G binding proteins and Escherichia coli intimin, are involved in bacterial pathogenesis (3). lytB (fivefold) encodes glucosaminidase, which catalyzes the last step in cell wall division that disperses intact cells (17). It has been shown recently that lytB mutants exhibited a significantly reduced colonization of the nasopharynx but exhibited a behavior similar to that of the parental lytB+ strain when they were tested in a model of pneumococcus-induced sepsis (22). A significant decrease of expression (10-fold) of spr2021 (pcsB) has been shown in microarray analysis and validated further by qRT-PCR. The pcsB homologues of Streptococcus mutans (11) and Streptococcus agalactiae (67) have been implicated in responses to general stress, but their functions remain unknown. pcsB mutants in both species of streptococcus were more sensitive to low pH, high osmotic pressure, and high temperature than was their pcsB+ parent strain (67, 68). A similar increase of sensitivity to elevated temperature, high osmolality, and pH was observed in an S. pneumoniae R6 derivative strain expressing less than half the normal PcsB amount (M. Winkler, personal communication). Unlike the case in S. mutans and S. agalactiae, PcsB is strictly essential in S. pneumoniae (51). It has been demonstrated that severe depletion of PcsB led to a rapid cessation of growth. Therefore, a model in which PcsB acts as a critical murein hydrolase that modulates cell wall biosynthesis has been proposed.
To test whether decreased resistance to stress conditions of the stkP mutant resulted from low abundance of pcsB transcript, we constructed a merodiploid strain expressing an ectopic copy of the pcsB allele under the control of a fucose-inducible PfcsK promoter. No significant effect on either growth characteristics or stress resistance was observed when the ectopic expression of pcsB was induced with fucose over a wide range of inducer concentrations (data not shown). Therefore, it seems that the increased sensitivity of the stkP mutant to stress conditions is multifactorial, not simply dependent on the decreased level of pcsB expression.
Recent studies revealed that pcsB expression is positively regulated by the VicRK two-component system (52). VicRK is a homologue of the essential two-component system, which is highly conserved in and specific to the low-G+C-content gram-positive bacteria, where it is often designated YycFG based on its initial discovery in Bacillus subtilis (20). Both the yycF and yycG genes are required for viability. Conversely, in S. pneumoniae, only the response regulator vicR gene was found to be essential (32, 74), suggesting that VicR functions independently of sensor kinase VicK, possibly participating in signal transduction with another two-component system kinase. In a recent study performed by Ng et al. (53), binding sequences for VicR were determined by band shift and footprinting experiments. VicR bound to regions upstream of pcsB, pspA, spr0096, spr1875, and spr0709. The phosphorylation of VicR to VicR-P increased the apparent strength and changed the nature of binding to these regions.
A comparison of our microarray results with those obtained by Ng et al. (52, 53) shows that spr0096, spr0709, lytB, and pcsB genes were strongly downregulated in both the vicR and stkP isogenic mutants. No changes were observed in the expression of the vicRKX genes in the stkP mutant compared to that in the stkP+ parent. Thus, the positive regulatory effect on the expression of spr0096, spr0709, lytB, and pcsB genes can be confined to direct interaction of StkP with VicR. This finding could suggest that in addition to its cognate kinase, VicK, the activity of the response regulator VicR can be modulated by StkP. Whether StkP influences the ability of VicR to bind DNA and/or regulate downstream transcription or affect the stability of VicR is currently being investigated. Recently, Rajagopal et al. (65) brought the first experimental evidence that direct cross talk between two different signaling systems exists.
The genes encoding proteins involved in pyrimidine de novo biosynthesis comprised the largest category, containing 10 genes that were coordinately downregulated by 2.4- to 5.2-fold. In contrast to a number of different gram-positive organisms, the pyrimidine biosynthesis genes are scattered on the chromosome of S. pneumoniae and are organized in five putative transcriptional units: pyrFpyrE, pyrDIIpyrD, pyrC, pyrP, and pyrRpyrBcarAcarB. The first gene in the pyrRpyrBcarAcarB operon is pyrR encoding a regulatory protein. In B. subtilis, regulation of the pyr operon occurs through a transcriptional attenuation mechanism in which PyrR promotes transcriptional termination at three attenuation regions in the operon when uridine nucleotide levels are high (34, 75). High levels of UMP stimulate PyrR to bind to a conserved sequence and secondary structure in the mRNA in each attenuation region called the antiantiterminator or the binding loop (73). Binding of PyrR to this sequence prevents the formation of an antiterminator stem-loop and allows a downstream 
-independent transcription terminator to form, reducing the expression of downstream genes. When UMP levels are low, the RNA binding affinity of PyrR is reduced and the more stable antiterminator stem-loop is favored, leading to transcriptional read-through and expression of the downstream genes.
By analyzing the sequences of the upstream and intergenic regions of genes comprising the PyrR regulon in S. pneumoniae, we identified the highly conserved sequence motif 5'-ARUCCNGNGAGGYU-3' in the pyrRpyrBcarAcarB mRNA leader, in the pyrBcarA intergenic region, and in the pyrFpyrE and pyrP leader sequences (data not shown). The secondary structures that may be formed by the RNA in these regions in S. pneumoniae are similar to the structures found in the RNA transcribed from the B. subtilis pyr operon. These findings suggest that the expression of the pyr genes of S. pneumoniae can be regulated by the same attenuator mechanism as that suggested for B. subtilis. The fact that all genes of the S. pneumoniae pyr regulon, including pyrR, are repressed in the stkP mutant could indicate that in addition to the attenuation mechanism, their expressions are regulated by an alternative mechanism. Collectively, these results suggest that the regulation of pyrimidine biosynthesis in S. pneumoniae involves multiple levels of control (i.e., transcriptional and posttranslational).
There were two- to fourfold decreases in the relative transcript amounts of several genes involved in DNA repair. Two of these, spr1055 and spr1054, coding for DNA uracil glycosylase (ung) and 8-oxodGTP nucleoside triphosphatase (mutX), respectively, are clustered with pyrC (see above). MutX of S. pneumoniae is a homologue of E. coli MutT, an antimutator protein (40). In addition, the expression of spr1317 (O6-methylguanine-DNA methyltransferase ogt) and spr1660 (exodeoxyribonuclease exoA) was moderately affected (twofold decrease).
Decreased expression of a gene cluster, piuBCDA (spr1684-spr1687), containing the components of an iron uptake ATP binding cassette (ABC) transporter (also called a fat or pit1 operon) that is required for virulence (6, 7) was observed (2.8- to 4.7-fold decrease). This result was further validated by qRT-PCR, which showed even higher values for the change in expression for the first gene in the operon, spr1684 (fatD) (Table 3). In addition, there was a 2.2-fold reduction in spr0934 transcripts, a homologue of piaA. piaA is the first gene of a four-gene iron transport operon (Pit2) and has been identified as a ferrophore binding protein (7). Three gene clusters, pit, pia, and piu, each encoding four elements of an iron uptake ABC transporter, have been implicated in iron uptake by S. pneumoniae (6, 7).
To test the susceptibility to inhibition by streptonigrin, which can be used as an indirect indicator of intracellular iron concentration (5), we measured the zone of growth inhibition surrounding an antibiotic disk impregnated with 5 µg of streptonigrin on blood agar plates. Since no differences in zone size were observed for mutant and wild-type strains, we concluded that both strains have similar iron transport capacities (data not shown). This result is in agreement with the previous finding that the pia product is likely to be the dominant S. pneumoniae iron transporter during both in vitro and in vivo growth (7). Thus, the attenuated virulence of the stkP mutant strain cannot be attributed to the defect in iron transport.
Another functional category of repressed genes comprises those that mediate resistance to oxidative stress, including spr1322 and spr1495, encoding pyridoxine biosynthesis protein Pdx1 and thioredoxin-linked thiol peroxidase PsaD. Pdx1 has been reported as a singlet oxygen resistance protein (19), and it has been suggested that PsaD is a homologue of E. coli thiol peroxidase, which scavenges superoxide and peroxide ions (54).
In addition to the genes with experimentally defined functions, a large group of repressed genes coding for proteins with either unknown or putative functions were observed. The alteration in relative transcription of several of these genes was considerable, suggesting that they may function in the observed phenotypic properties. A notably large decrease for the spr1057-spr1056 cluster (59- and 9-fold, respectively) encoding conserved hypothetical polypeptides was detected. spr1057, the most strongly repressed gene (as high as 500-fold by qRT-PCR) in the stkP mutant, encodes a homologue of the haloacid dehalogenase (HAD)-like protein superfamily, which includes a variety of enzymes with different functions. Despite the fact that it is named after a dehalogenase, the vast majority of the known catalytic activities in the HAD family involve phosphoryl transfer (2). The few members of the family that have a defined function are associated with membrane transport, metabolism, signal transduction, and nucleic acid repair. The E. coli YjjG protein, a member of the HAD superfamily, exhibits a high phosphatase activity towards 5'-dTMP, 5'-dUMP, and 5'-UMP and might be involved in the pyrimidine substrate cycles (63). Further work is necessary to elucidate the function of the HAD protein in S. pneumoniae. Nevertheless, the extent and strength of its repression in the stkP mutant suggest that its expression can be completely inhibited. Indeed, a proteomic analysis of the wild-type and stkP strains demonstrated that mutant cells are depleted of Spr1057 protein (data not shown).
Among the other repressed hypothetical genes, we observed a substantial reduction (10-fold by microarrays and 15-fold by qRT-PCR) in the expression of the spr1320 gene encoding a putative second pneumococcal hemolysin that is distinct from pneumolysin. Spr1320 is a conserved protein in all streptococci as well as additional bacterial species and is similar to hemolysin III. It has been described as a lysin responsible for a beta-hemolytic effect in pneumococci (9); however, no data are available regarding its potential role in pathogenesis.
Inactivation of stkP leads to overexpression of competence genes.
The analysis of the induced genes in the stkP mutant showed that the majority of upregulated genes belong to the competence regulon (Table 2). This finding was surprising as it has been previously demonstrated that an stkP mutation effectively abolishes natural competence development (18). Furthermore, the S. pneumoniae strains were grown in CAT medium, in which the wild-type strain CP1015 fails to develop spontaneous competence. We selected these culture conditions to uncouple the regulatory effects of StkP on growth and physiology from those on competence. All the ComE-dependent genes (early com genes) described previously as key components of the competence regulatory cascade, including comCDE, comAB, comX, and comW (26, 33, 35, 36, 58), were strongly upregulated in the stkP mutant (Table 2). These genes encode functions absolutely required for competence induction. A comparison of the induction ratios of the core early competence genes, including comCDE, comAB, comX, and comW, with those reported in previous reports (15, 61) suggested that the competence regulatory circuit is fully induced. Although there were differences in the amplitude of induction between particular surveys, these differences could be attributed to the distinct microarrays, variations in medium composition, and in the strains used.
In contrast, the levels of induction of 
x-dependent genes (late competence genes) required for the process of DNA uptake, processing, and recombination were generally 10- to 20-fold lower than those reported previously (15, 61). In addition, several competence genes had transcript levels below the cutoff value of twofold, suggesting that their expression levels were not affected in the stkP mutant. These genes include celA, coiA, dalA, cclA, recA, cglC, cflA, and cflB. It has been experimentally demonstrated that these genes encode proteins whose functions are indispensable in efficient DNA binding, uptake, and recombination (8, 38, 46, 59, 61). Among these, the absence of recA induction was confirmed by qRT-PCR (Table 3). Previous studies have demonstrated that the recA gene is induced at competence (38, 57), and de novo synthesis of RecA molecules is absolutely required for full recombination proficiency during transformation (46, 61). recA is the second gene in a cinA-recA-dinF-lytA late operon and, in competent cells, is induced from the competence-inducible promoter located in front of cinA (Pcin). In noncompetent cells, recA is expressed from its own A-type promoter (46). The competence-specific induction of recA (that is, expression from Pcin) accounts for 95% of the yield of transformants. Thus, we conclude that low basal level expression of recA and possibly other DNA uptake and processing genes could be a major factor contributing to the reduced rate of transformation in the stkP mutant strain.
Responsiveness of competence genes to CSP induction in the wild-type and stkP strains.
The results from our global transcription analysis showed that the "core" early competence genes are induced in the stkP mutant during mid-exponential growth. In order to address whether an observed increase of expression of selected competence genes in the stkP mutant corresponds to the levels of the fully induced com regulon, we quantified the mRNA levels in the CSP-induced and uninduced cultures of both wild-type and mutant strains. comC, comE, comA, comX, and comW were selected as representatives of early genes. Cultures of both strains were grown in parallel to an OD400 of 0.4 (the same conditions as cultures used for microarray analysis) and sampled immediately before and after the addition of CSP (250 ng ml–1) for the purification of RNA (see Materials and Methods). This dose of CSP was sufficient to induce full competence in the wild-type strain. To analyze the kinetics of the CSP response, samples were taken 8 and 13 min after CSP addition. As expected, in the wild-type strain, a low basal signal increase was observed for representatives of early genes (comC, comE, comA, comX, and comW) tested upon CSP treatment, reaching the maximum level at 8 min and then declining by 13 min (Fig. 3). This rapid response and sharp decline in mRNA levels are typical of what was previously reported for competence genes (1, 60, 69). In contrast, the level of the corresponding transcripts in the mutant strain, although significantly elevated, remained unaffected by the addition of CSP (Fig. 3). Transcription levels of the induced genes in the wild-type strain reached values two- to eightfold higher than those measured in the stkP mutant.
|
Competence development.
To exclude the possibility that the observed failure of stkP strain cells to respond to exogenous CSP may result from an occurrence of a transient postcompetence period, when cells remain refractory to CSP (43), we performed the analysis of transcript levels as well as transformation assays at a lower optical density (OD400 = 0.1). This density was selected because the refractory phase persists for about 1 to 1.5 generations after spontaneous as well as CSP-induced development of competence (12, 43). The cell counts at given optical densities corresponded to approximately 1.5 generations. The transcript level of the sampled genes was essentially the same as that in the cultures at a higher cell density, suggesting that the absence of a CSP response in the mutant strain is not a consequence of a refractory period (data not shown). Competence assays were performed with cultures of the wild-type and stkP mutant strains. Competence development was measured in cultures both induced by the addition of exogenous CSP and uninduced and grown in CAT medium, a broth which prevents Cp1015 from developing spontaneous competence. As expected, basal transformation efficiency was observed in the wild-type cultures in conditions that repressed competence development (<10–5%) (Table 4). In the isogenic stkP mutant strain, the yield of transformants attained a transformation efficiency 
1,000-fold higher than that of the wild type. The addition of CSP to the wild-type culture resulted in a 105-fold increase in the transformation frequency (it reached 2.3%), whereas no change was observed in the stkP mutant strain. This result demonstrates that the transformation efficiency in the mutant strain is not affected by the addition of exogenous CSP. The transformation efficiencies of CSP-treated and untreated cells were similar to those obtained at an OD400 of 0.4 (Table 4). Thus, it seems plausible that the competence regulon is permanently induced; however, the degree of induction does not reach sufficient levels for full activation of the competence cascade. Thus, the results of the transformation assays are consistent with those of qRT-PCR and microarray analysis.
|
In summary, our microarray analysis of the transcription profile of a stkP mutant revealed that the posttranslational modification of an as-yet-unidentified target(s) affects the expression of many functionally different genes in S. pneumoniae. Nevertheless, considering the fact that the contribution of stkP to virulence varies with the different S. pneumoniae strains and the site of infection (18), it is very likely that the impact on the repertoire of genes regulated will be strain dependent. It is becoming increasingly clear that genes encoding signaling proteins make different contributions to gene expression depending on the genetic background of the parental strain (25).
The broad pleiotropic effect of the stkP mutation is noteworthy. One possible explanation is that some phenotypic properties of the mutant strain can be ascribed to the deregulation of an essential GlmM enzyme (55), resulting in so-far-unidentified, although envisaged, alterations in the cell wall biosynthesis. These perturbations could be reflected in some features of the mutant strain likely involving susceptibility to stress conditions. Apart from this possibility, StkP could assume a direct role, interacting with transcription machinery. However, as StkP lacks canonical DNA binding domains and as a result is unlikely to directly regulate transcription, the input signal must be transmitted through an effector molecule. Considering that this effect is exerted to many different gene categories involved in diverse cellular processes, the existence of a common factor mediating the transfer of signal to transcription machinery seems plausible. We hypothesize that the -subunit of RNA polymerase is a candidate for such an effector. Our studies revealed that RpoA is a substrate of StkP either in vivo (55) or in vitro (unpublished results). Its covalent modification, particularly at the C-terminal domain (23), can contribute to the altered efficiency of transcription initiation by changing its affinity for promoters or transcription factors (70). The identification of the target amino acids that are phosphorylated will provide the grounds for further investigation.
|
ACKNOWLEDGMENTS |
|---|
We thank Donald Morrison for valuable discussions. We also thank Malcolm Winkler for provision of the pcsB-complementing amplicon and Vladimír Ko
ínek for help with qRT-PCR. We are grateful to Elaine Allan for critically reading the manuscript.
|
FOOTNOTES |
|---|
ská 1083, 142 20 Prague 4, Czech Republic. Phone: (42 02) 41 06 26 58. Fax: (42 02) 41 72 22 57. E-mail: branny{at}biomed.cas.cz 
Published ahead of print on 6 April 2007.
Supplemental material for this article may be found at http://jb.asm.org/.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
