![[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}
Departments of Pharmacology,1 Medicine, University of Wisconsin Medical School,2 School of Veterinary Medicine, University of Wisconsin College of Agriculture and Life Sciences, Madison, Wisconsin3
Received 5 May 2004/ Accepted 29 July 2004
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Genome sequencing in Streptococcus pneumoniae revealed 13 TCST pairs along with one orphan response regulator (26, 52, 53). Throup et al. (53) inactivated all 14 response regulators by antibiotic cassette mutagenesis in a type 3 encapsulated strain and demonstrated that 9 of these response regulators were required for murine respiratory tract infection, including the orphan response regulator gene designated rr489 (referred to here as ritR). The ritR locus is not adjacent to a kinase gene, as the other 13 TCST kinase-RR pairs are, and RitR is not known to be phosphorylated by any of the 13 known histidine kinases.
Iron serves as a major factor in microbial virulence both as an essential element and as an element that is able to catalyze the synthesis of reactive oxygen intermediates (ROIs) from H2O2, commonly referred to as the Fenton reaction. The interaction between Fe2+ and H2O2 can be particularly deleterious under iron overload conditions (for reviews, see references 2 and 24). Despite the widespread abundance of iron in nature, the free Fe3+ concentration is estimated to be in the range of 10–18 M at mucosal surfaces (9). Sequestration of iron by host iron-binding proteins, such as hemoglobin, lactoferrin, and transferrin, contributes to its low concentration. In response, bacteria have evolved elaborate mechanisms to capture iron needed for growth while minimizing iron-based toxicity.
The iron-binding transcriptional regulator protein Fur has been shown to regulate iron uptake in both gram-positive and gram-negative bacteria. Fur binds intracellular free Fe2+ directly, increasing its affinity for a conserved DNA motif, the Fur box, a regulatory sequence associated with iron uptake (16). Fur regulation also plays a role in the oxidative stress response (39) and the regulation of other iron-related genes (for reviews, see references 2 and 19). Gram-positive bacteria with a high G+C content, such as corynebacteria and mycobacteria, utilize iron uptake repressors belonging to the DtxR family, which has little or no homology to Fur. In Corynebacterium diphtheriae, DtxR regulates siderophore biosynthesis, iron uptake, and toxin production (51). A potential DtxR family member, SmrB, which resembles TroR (40), a regulator of Mn uptake in Treponema pallidum, has been identified in S. pneumoniae (27); however, a role for this protein in iron homeostasis has not been established.
Three gene clusters, pit, pia, and piu, each specifying a set of four elements of an iron uptake ATP-binding cassette (ABC) transporter, have been implicated in iron uptake by S. pneumoniae (5, 6). In the studies reported here, DNA microarrays were used to learn which genes RitR might regulate. Arrays revealed that RitR was associated with the transcription of genes involved in iron uptake and the oxidative stress response. In DNA footprinting experiments, recombinant glutathione S-transferase (GST)-RitR was shown to bind directly to three sites in the promoter region of the piu operon, thereby providing a direct link between RitR and iron uptake regulation. We also show that ritR is cotranscribed with gnd (encoding 6-phosphogluconate dehydrogenase), suggesting that there is coordinated regulation of iron homeostasis and the pentose phosphate pathway.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
F' and RR1 were grown in Luria-Bertani medium at 37°C overnight with aeration. The antibiotic concentrations used for plasmid selection in S. pneumoniae were as follows: chloramphenicol, 2 µg/ml; and erythromycin, 1 µg/ml. The antibiotic concentrations used for plasmid selection in E. coli were 34 µg of chloramphenicol per ml and 500 µg of erythromycin per ml. S. pneumoniae strains were transformed as described previously (20).
|
|
Preparation and hybridization of cDNA probes. RNA was obtained from S. pneumoniae cells grown in 100 ml of BHI liquid medium to an A600 of 0.35 to 0.40, as described by de Sazieu et al. (13). For analysis of transcripts under tet control, 0, 100, and 300 ng of anhydrotetracycline per ml were added at an A600 of approximately 0.2, and cells were grown to an A600 of 0.35 to 0.4. RNA was harvested for cDNA synthesis. cDNA was synthesized by using 5 U of SuperScriptII reverse transcriptase (Invitrogen) in a 100-µl reaction mixture with 100 µg of total RNA as the template. Reaction mixtures were incubated for 20 h at 42°C. To degrade RNA, 14 µl of 1 N NaOH was added, and the reaction mixtures were heated for 15 min at 65°C. The reaction mixture was neutralized with 14 µl of 1 N HCl. The cDNA preparation was purified and concentrated by using a Microcon-30 microfiltration device (Amicon, Bedford, Mass.) and dissolved in 35 µl of water. The cDNA was then partially digested by addition of 12 U of DNase I (Promega), after a pilot incubation to determine the time needed to optimize the yield of 50- to 100-bp fragments for microarray hybridization. The DNase I was denatured at 99°C for 15 min to stop the reactions, and the resultant preparation was washed three times with a Microcon-10 centrifuge column. Digested cDNA was then collected in 30 µl of water and 3' end labeled with 2 nM biotin-6-ddATP (NEN) and 50 U of deoxynucleotidyl terminal transferase (Promega).
Probes were hybridized to the arrays by using a protocol that specified the use of reagents and laboratory equipment provided by NimbleGen Systems Inc. (Madison, Wis.), as follows. All procedures were performed at 42°C unless otherwise noted. Arrays were loaded onto a hybridization wheel (NimbleGen Systems) in individual sealed chambers and incubated with prehybridization buffer (PHB) (100 µg of salmon sperm DNA per ml, 50 µg of acetylated bovine serum albumin [BSA] per ml, 2x morpholineethanesulfonic acid [MES] buffer) for 15 min and then with hybridization buffer (HB) (PHB plus 3 to 15 µg of biotin-labeled cDNA, 1 nM nonspecific competitor oligonucleotide) for 15 min. Before PHB and HB were loaded into hybridization chambers, the buffers were incubated at 98°C for 5 min, centrifuged at 14,000 rpm for 15 min using an Eppendorf model 5414 microcentrifuge, and kept at 45°C until use. Four hundred microliters of PHB was added to each array in the hybridization wheel and incubated at 42°C for 15 min, and then 400-µl portions of the HB samples were added.
Arrays were hybridized overnight at 42°C on a rotating platform for 16 to 20 h. Arrays were then washed three times with nonstringent wash buffer (NSWB) (6x SSPE, 0.01% Tween 20) (1x SSPE is 0.18 M NaCl, 10 mM NaH2PO4, and 1 mM EDTA [pH 7.7]) and then five times with stringent wash buffer (0.1 M MES buffer [pH 6.5], 26 mM NaCl, 0.01% Tween 20), and this was followed by staining with Cy3-streptavidin. Subsequent washes were performed at the ambient temperature unless otherwise noted. Primary stain buffer (Cy3-streptavidin, 5 mg of acetylated BSA per ml) was then added, and arrays were incubated for 15 min at room temperature. Arrays were then washed three times with 400 µl of NSWB, which was followed by addition of 400 µl of secondary stain buffer (5 µg of biotinylated immunoglobulin G per ml, 100 µg of goat immunoglobulin G per ml) and incubation for 15 min. Following three additional washes with NSWB, slides were briefly washed in 50 ml of NSWB with agitation and then with NimbleGen final wash buffer (proprietary) for 30 s. Arrays were then dried briefly with argon and scanned with an Axon Genepix 4000B scanner at 532 nm and a resolution of 5 µm.
High-density oligonucleotide microarray.
The genome sequence and open reading frame (ORF) predictions for the S. pneumoniae R6 genome were obtained from GenBank accession number NC-003098. The genome sequence file contains 2,038,615 bases and encodes 2,043 predicted open reading frames. A high-density oligonucleotide array for the S. pneumoniae genome was designed for this study containing 40,860 probe pairs, where each pair contains a perfect match (PM) oligonucleotide and a mismatch (MM) oligonucleotide, for a total of 80,142 different 24-nucleotide array feature sequences. The PM sequence is a 24-mer oligonucleotide probe from the S. pneumoniae genome. MM probes were identical except for substitutions at positions 6 and 12 from the 5' end. To generate MM sequences, bases were systematically substituted by inversion as follows: A 
T, C G, G C, and T A. By using the NimbleGen probe assessment and selection software (NimbleGen Systems), potential probe sequences were checked for internal complementarity that would reduce their ability to hybridize optimally. Oligonucleotide features were also checked for sequence redundancy by comparing each probe to the entire S. pneumoniae genome sequence. Twenty distinct high-quality probe pairs were selected to represent each ORF.
Microarray data analysis methods.
Following hybridization, the arrays were scanned, and the median signal intensity for each probe on the array was calculated by using NimbleGen's extraction software. For each probe pair the difference between the PM and MM signal intensities was calculated together with the Tukey biweight mean (21) (biweight constant = 6) from the 20 probe pairs for each ORF. For purposes of normalization of the array intensity data, array-specific scaling factors were calculated by assuming a constant mean signal intensity of 1,000 signal units for each array. The signal for each ORF was obtained by multiplying the raw signal intensity by the array-specific scaling factor. The program Ebarrays (25) was used to identify ORFs with significant changes in expression comparisons of 
ritR versus ritR+ or ritR versus tet-induced expression of ritR. The initial parameter settings were as follows: p1 = 0.95, p2 = 0.05, theta1 = 16, theta2 = 1, theta3 = 25. ORFs were selected as significant if the posterior probability associated with differential expression was 
0.5.
Mouse infection studies.
S. pneumoniae strains R800 (ritR+) and RU402 (ritR) were used in thigh and lung infection models as described by Andes and Craig (1). Cyclophosphamide-treated mice were inoculated with 4 x 108 cells intranasally or by injection into the thigh. Organs were harvested at zero time and 48 h and homogenized, and bacterial counts were determined by plating homogenate dilutions on TSBA. The data reported below are the averages of counts obtained from three mice for each time.
ritR-regulated expression of LacZ. LacZ activity was measured fluorimetrically as described previously (55). S. pneumoniae strain CP401 was grown overnight in CAT medium and diluted 1:5 in 20 ml of fresh BHI medium. One-milliliter samples were collected by centrifugation during growth of the culture, and the resultant cell pellets were washed, resuspended in 100 µl of phosphate-buffered saline containing 1% Triton X-100, and transferred to a 96-well microtiter plate. The test reaction mixture was supplemented with 4 µg of methylumbelliferyl galactoside and incubated for 5 to 10 min at the ambient temperature. Fluorescence was excited at 335 nm, and emitted fluorescence was read at 480 nm with a microtiter plate reader (Titertek Fluoroskan).
Northern blot analysis. Primers 5F, 5R through 12F, and 12R (Table 2) were used to amplify probes for Northern blot analysis. S. pneumoniae total RNA was prepared as described above and analyzed by using standard methods for electrophoresis, transfer, probing, and detection by autoradiography or by use of a phosphorimager (43).
Iron-related toxicity. Hydrogen peroxide toxicity was determined as described by Tseng et al. (54), and streptonigrin toxicity was determined as described by Brown et al. (5). Cells were plated on TSBA to measure the surviving cell titer following exposure to the agent.
Gene nomenclature. The different gene and locus names associated with the iron transporter subunits mentioned in this paper are cross-referenced in Table 3.
|
GST-RitR-piuB gel mobility shift.
The piu regulatory region from coordinate –107 to coordinate 110 (217 bp) relative to the transcriptional start site of piuBCDA was amplified by PCR by using primers 13F and 13R. The resultant product was cleaned with a QIAquick PCR purification kit (QIAGEN) and 5' end labeled in a 25-µl (total volume) mixture with [
-32P]ATP (specific activity, >6,000 Ci/mmol; 20 µCi) by using 20 U of phage T4 polynucleotide kinase (New England Biolabs). Unincorporated nucleotides were removed by passage through a G50 spin column (Pharmacia). The gel shift reaction mixtures (total volume, 10 µl) contained 1x binding buffer (20 mM HEPES [pH 7.2], 5 mM MgCl2, 1 mM CaCl2, 0.1 mM EDTA, 10% glycerol) supplemented with 10 ng of -32P-labeled piuB promoter, 800 ng of poly(dI-dC), and GST-RitR at a series of final concentrations that ranged from 0 to 8 µM. The reaction mixtures were incubated at room temperature for 10 min and loaded onto a 1x TAE (10 mM Tris-acetate [pH 8.0], 1 mM EDTA) nondenaturing 4% polyacrylamide gel, which was run at 16 V/cm (constant voltage) for 1 h. Fractionated RitR-DNA complexes were visualized by autoradiography.
GST-RitR footprint analysis of the piuB promoter region.
PCR primers 13F and 13R were 5' end labeled in a 25-µl (total volume) mixture with [-32P]ATP (specific activity, >6,000 Ci/mmol; 20 µCi) by using 20 U of phage T4 polynucleotide kinase. The labeled primers were cleaned with a QIAquick PCR purification kit (QIAGEN), and unincorporated nucleotides were removed by passage through a G50 spin column (Pharmacia). The piu regulatory region from coordinate –107 to coordinate 110 (217 bp) relative to the transcriptional start site of piuBCDA was amplified by PCR in two reaction mixtures. One mixture contained labeled primer 13F plus unlabeled primer 13R, and the other contained unlabeled primer 13F plus labeled primer 13R. The two PCR products generated were used to produce forward and reverse DNA footprints, respectively. Two picomoles of reverse or forward -32P-labeled piu promoter was incubated in 50-µl (total volume) mixtures containing 1x binding buffer supplemented with 1 mM dithiothreitol, 800 ng of dIdC, and a series of GST-RitR concentrations which ranged from 0 to 30 µM. The reaction mixtures were incubated at room temperature for 10 min, after which 0.6 U of DNase RQ1 (Promega), diluted 1:15 from a 1-U/µl stock solution with dilution buffer (Promega) containing 20 mM Tris-HCl (pH 7.0) and 5 mM MgCl2, was added. After 1 min of incubation, the reactions were quenched with stop buffer (200 mM NaCl, 30 mM EDTA, 1% sodium dodecyl sulfate), and the preparation was purified by extraction twice with an equal volume of phenol. The resultant preparation was precipitated with ethanol and resuspended in 4 µl of loading buffer (0.5x Tris-borate-EDTA, 0.05% bromophenol blue, 0.05% xylene cyanol, 8 M urea). Samples were heated at 95°C for 40 s before they were loaded onto a 10% acrylamide denaturing gel in 0.5x Tris-borate-EDTA-urea. A Maxam-Gilbert sequencing ladder was run adjacent to the samples generated from the same PCR fragment that was used in the footprint reactions (46).
Reverse transcriptase mapping of the piuB promoter transcription start site.
The 65-µl primer extension annealing reaction mixture contained 30 µg of RU402 purified total RNA, 1 pmol of primer 13R labeled as described above, and each deoxynucleoside triphosphate at a concentration of 20 mM. The reaction mixture was heated to 95°C for 5 min and cooled to 42°C, after which it was supplemented with 35 µl containing 3 U of SuperScript II reverse transcriptase (Invitrogen), 10 µl of 0.1 M dithiothreitol, and 20 µl of 5x reverse transcriptase buffer (final volume, 100 µl). The resultant reaction mixture was incubated at 42°C for 2 h. A sequencing ladder was synthesized by using a DNA sequencing kit (USB Sequenase, version 2.0) according to the manufacturer's specifications. The 10-µl annealing reaction mixture contained 1 pmol of PCR primer 13R (unlabeled) and 1 pmol of the piuB promoter region. The annealing mixture was brought to 100°C for 5 min and then cooled on ice for 5 min. Sequencing reactions were initiated with addition of 1 µCi of [-33P]ATP (specific activity, 3,000 Ci/mmol) to each of the four termination reaction mixtures (one containing ddATP, one containing ddGTP, one containing ddCTP, and one containing ddTTP). Electrophoresis and autoradiography were conducted as described above for the footprint analysis. The promoter start site and sequencing ladder reaction mixtures were fractionated in parallel and visualized by autoradiography.
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To validate the use of R800 for the present studies, the copy of ritR present in R800 was inactivated by using the insertion-deletion vector pAU402. The resultant strain, RU402, was tested for its ability to infect mouse lungs and thighs as described by Andes and Craig (1). Although R800 is rough and therefore less pathogenic than the type 3 encapsulated strain, its pathogenicity can by enhanced by infecting test mice that have been immunosuppressed by treatment with cyclophosphamide (1). The results (Fig. 1) indicate that inactivation of ritR reduced the ability of S. pneumoniae to proliferate in the mouse lung but did not reduce its ability to infect the thigh. In contrast, R800 (ritR+) cells were able to proliferate in both tissues.
|
ritR mRNA levels are maximal during exponential growth in vitro. The timing of ritR expression during the growth cycle of S. pneumoniae indicates when to harvest cells for cDNA microarray probe preparation. Plasmid pEVP3 has been used for insertion-deletion gene expression studies (38). It contains a promoterless E. coli lacZ gene and can be integrated into a selected location by single-crossover site-specific recombination if it is provided with a sequence at the desired integration site. A 353-bp fragment of ritR was ligated into pEVP3 at its BamHI site. Following transformation into E. coli, constructs containing the ritR insert in both the forward and reverse directions were obtained, yielding plasmids pAU401 and pAU402, respectively.
The recipient S. pneumoniae strain, CP1250, is deficient in ß-galactosidase activity and can therefore be used in experiments that monitor the activity of E. coli lacZ as a reporter (38). Strain CPU401 was obtained by transformation of the pAU401 construct into CP1250. In addition, the control strain CPU402 was obtained by incorporation of the pAU402 construct into CP1250, integrated in the opposite orientation with respect to pAU401. Integrated constructs were selected on the basis of chloramphenicol resistance, and integration at the expected site in the chromosome was verified by DNA sequencing. Strain CPU401 was grown in BHI liquid medium and sampled, as shown in Fig. 2. LacZ activity in the cell pellets was measured fluorimetrically by using methylumbelliferyl galactoside as the substrate. The results shown in Fig. 2 indicate that LacZ specific activity in the culture increased sharply during early exponential growth until the A600 was 0.3 and remained relatively constant until the onset of the stationary phase at an A600 of about 1.1, whereupon the LacZ specific activity decreased sharply. RNA for microarray studies was therefore harvested in the early exponential phase of growth at an A600 of 0.35 to 0.40, when ritR mRNA and RitR protein would be expected to be maximally abundant.
|
ritR and ritR+ strains.
Nothing is known about the regulatory targets of RitR. As an initial approach to this problem, the transcriptional profile of S. pneumoniae RU402 (ritR) and that of its parent strain, R800 (ritR+), were compared by using high-density DNA microarrays. Microarrays based on the S. pneumoniae R6 DNA sequence (23) were chemically synthesized as described by Nuwaysir et al. (33). The arrays were probed with cDNA obtained by reverse transcription of total RNA harvested during the early log phase of growth. Array data from three separate experiments were normalized and averaged as described in Materials and Methods.
The statistical program EBarrays, developed by Kendziorski et al. (25), was used to determine if the two gene expression profiles of R800 and RU402 differed from each other in a statistically significant manner. The basis for this determination is the EBarrays-derived value called the probability of differential expression (PDE). In this study a gene was considered differentially expressed if (i) the ritR/ritR+ expression profile average log2 ratio was 1.0 or 
–1.0 and (ii) its calculated EBarrys PDE value was 0.5. By applying these criteria, 54 genes were found to differ significantly in the ritR-ritR+comparison; 17 genes were repressed, and 37 were activated. In this report up-regulated and down-regulated refer to the expression of a given gene in the wild-type R800 strain (ritR+). Results are shown in Table 4. A scatter plot of genes that were differentially expressed in the ritR-ritR+ comparison is shown in Fig. 3. Red squares indicate genes which had PDE values of <0.5, and blue squares indicate genes with PDE values of 0.5.
|
|
mRNA levels activated by ritR.
There were the following noteworthy changes in mRNA levels that were activated by ritR. (i) ritR had the highest average log2 ratio, 4.0, indicating that it had indeed been activated in R800 (ritR+) relative to RU402 (ritR). These observations confirmed that ritR had been effectively knocked out in RU402. (ii) There were changes in homologs of Dpr, an iron storage-peroxide resistance protein, and in AdhE, an iron-binding alcohol dehydrogenase, both of which are implicated in H2O2 resistance in E. coli (15, 42) (average log2 ratios, 2.51 and 1.8, respectively). Also activated were (iii) HemH, which is responsible for catalyzing the addition of Fe2+ to the protoporphyrin ring, the final step in hemin synthesis (average log2 ratio, 2.18) (for a review, see reference 34); (iv) MutY, an iron-sulfur cluster binding A/G glycosylase used in DNA A/G mismatch repair (17) (average log2 ratio, 2.15); and (v) the loci spr1620 and spr1527, coding for two sugar-binding proteins (average log2 ratios, 3.31 and 1.89, respectively), and the locus spr0750, coding for a membrane-spanning permease component of a putative ribose/galactose ABC sugar transporter (average log2 ratio, 2.14). These data suggest that ritR is linked to the activation of genes required for the oxidative stress response, heme biosynthesis, and sugar uptake.
Comparative transcriptome analysis of gene expression by ritR under tet control. To measure changes in mRNA levels as a function of ritR expression, ritR was placed under control of an inducible tet promoter. Anhydrotetracycline was used to activate transcription over a range of concentrations that produced ritR transcript levels that varied from undetectable to maximally expressed. The 353-nucleotide fragment of ritR used previously to construct plasmids pAU401 and pAU402 was cloned into the BamHI site of plasmids pRK01 and pRK02. pRK01 contains a weak low-background tetracycline-inducible promoter, P57, whereas pRK02 contains the stronger P57-opt promoter and can be used to induce transcription levels severalfold compared with those obtainable with pRK01 (47). The resultant pRK01 and pRK02 constructs, designated pAU403 and pAU404, respectively, were integrated into the S. pneumoniae R800 chromosome following selection on TSBA containing erythromycin to obtain strains RU403 and RU404. RU403 and RU404 were exposed during the early log phase to 0, 100, and 300 ng of anhydrotetracycline per ml for 1 h before total RNA was harvested. The resultant RNA preparations were reverse transcribed and hybridized to high-density DNA microarrays as described above.
Due to the variability of differentially expressed genes, possibly as a result of the degree of RitR activation at different cellular concentrations, the data for the six microarrays that were probed with cDNA from tet-induced strains RU403 and RU404 were averaged and compared to the mutant RU402 (ritR) probed array data. PDE values were assigned to the averaged and sorted log2 ratios for comparison. Of 290 genes with PDE values of 0.5 and log2 average ratios of 1 or –1, 245 were found to be activated in R800 (ritR+), and 45 were repressed. A partial list corresponding to the 54 entries (of the 290 genes) that were also present in the RU402-versus-R800 comparison is shown in Table 4. The complete list of 290 entries for the tet induction series can be accessed at http://weisblumlab.pharmacology.wisc.edu.
The results from the array hybridizations based on tet-induced ritR expression paralleled those from the ritR-probed arrays. Most PDE values and average log2 ratios of activated genes from both the RU402-versus-R800 and tet-induced ritR array experiments correlated; there were two notable exceptions, spr0294, a phosphotransferase system sugar-specific EII component, and spr1620, a sugar-binding protein-ABC transporter component (Table 4). In contrast, genes that were repressed in the RU402-versus-R800 comparison varied considerably with respect to both the PDE and log2 ratio values of the tet-induced ritR array experiments. Notable exceptions were piuB, piuA, ugd, rr479, and the genes encoding several hypothetical proteins. Interestingly, some genes, such as rr479, were activated in the tet-inducible cDNA-probed arrays but were previously repressed in the RU402-versus-R800 comparison. The difference observed might be explained by the absence of regulation by RitR at the nonphysiological levels attainable in RU403 and RU404 cells carrying the tet-inducible ritR construct.
Linkage between ritR and the pentose phosphate pathway. The immediate upstream and downstream neighbors of ritR in the S. pneumoniae chromosome are gnd (encoding 6-phosphoguconate dehydrogenase) and cpbF (encoding choline-binding protein F), respectively. Northern blot analysis was used to test for transcriptional linkage in the gnd-ritR-cbpF region. 5'-32P-labeled probes specific for ritR, gnd, and cpbF were used to probe Northern blots of RNA harvested from R800 cells. Results shown in Fig. 4 indicate that the gnd and ritR probes hybridized to a 2.2-kb fragment, whereas the cpbF probe hybridized to a shorter 1-kb fragment. Since 2.2 kb correlates with the predicted length of the combined gnd and ritR ORFs, we inferred that ritR is transcriptionally linked to the upstream gene gnd but not to the downstream gene cpbF, which is transcribed independently. Transcriptional activation of the gnd-ritR gene cluster would therefore up-regulate gnd, which encodes a limiting enzyme of the pentose phosphate pathway.
|
Northern blot analysis of regulation by ritR.
To verify some of the key microarray observations, DNA probes representing piuB, piuA, dpr, ritR, and vncS were PCR amplified, 5' end labeled with -32P, and analyzed by Northern blot hybridization to fractionated total RNA isolated from RU402 and R800 (Fig. 5A). Also included were RNA samples from the tet-induced RU403 and RU404 strains grown in the presence of 0, 100, or 300 ng of anhydrotetracycline per ml (Fig. 5B). Results shown in Fig. 5A indicate that the level of ritR mRNA in RU402 cells was undetectable, whereas the piuB and piuA mRNA levels were detectable only in RU402 cells; the dpr mRNA levels were reciprocally regulated relative to piuB. Figure 5B confirms that the expression of ritR was induced by anhydrotetracycline compared to wild-type levels and that the piuA mRNA levels decreased as the RitR levels increased. Moreover, Fig. 5B indicates that the vncS mRNA levels paralleled those of induced expression of ritR, despite its PDE value, 0.2 (log2 ratio, 1.3), in our tet-induced ritR array data, indicating a possible linkage between the expression of these two genes. This observation is in contrast to the RU402-versus-R800 array data, where vncS exhibited a log2 ratio of –1.86, indicating repression by RitR. This could have been due to differences in the phosphorylation state of RitR in the two sets of arrays, with RitR insufficiently regulated at higher levels in the tetracycline-induced arrays compared with a wild-type level of regulation in the RU402-versus-R800 array experiments. Consistent with the findings of Brown et al. (5), Fig. 5C shows that piuB and piuA Northern probes both hybridized to a 4-kb fragment, which is the predicted size of the piuBCDA polycistonic mRNA, confirming that these genes are indeed cotranscribed as part of a four-gene operon. This observation contrasts with the apparent absence of detectable differential expression of the two internal genes in the piuBCDA operon, piuC and piuD, in the array studies. Site-specific processing or degradation of the piu polycistronic mRNA or a lack of sufficient cDNA for these two genes may account for this observed inconsistency.
|
ritR) due to the derepression of the piu operon. To address this question, RU402 and R800 cells were grown in iron-depleted medium supplemented with Fe2+, Fe3+, hemin, or hemoglobin, and susceptibility to streptonigrin was tested. Results shown in Fig. 6A and B indicate that there was little difference in susceptibility to streptonigrin between RU402 and R800 in deferrated medium. In contrast, supplementing the deferrated medium with exogenous Fe2+ and with hemin reduced the viability of RU402 relative to that of R800 by 5 and 3 orders of magnitude, respectively. In contrast, supplementation of the medium with Fe3+ or hemoglobin, which S. pneumoniae is also able to use as sources of iron (49), resulted in no observable difference in survival. These data imply that ritR represses the acquisition of hemin and free Fe2+ in vitro and not the acquisition of hemoglobin or Fe3+.
|
To address this question, S. pneumoniae R800 (ritR+) and RU402 (ritR) cells were challenged with 40 mM H2O2 as described by Tseng et al. (54). Each treated culture was sampled after exposure to H2O2, as indicated in Fig. 6C, and surviving cells were counted by plating on TSBA. The resultant killing curves, shown in Fig. 6C, indicate that R800 cells survived 10 min longer than RU402 cells under the conditions that were used. In contrast, use of paraquat at a concentration of 60 mM resulted in no observable differences (data not shown). Consistent with this observation, the microarray data showed that S. pneumoniae Fe and Mn superoxide dismutases were not differentially expressed in either set of array experiments (data not shown), indicating that ritR affects genes that either reduce Fe-catalyzed synthesis of ROIs or remediate their genotoxic effects but do not affect gene regulation or detoxification associated with superoxide.
GST-RitR-piu promoter reverse transcriptase mapping and gel mobility shift. Total RNA from RU402 enriched for piu mRNA (due to derepression of piu) was used to determine the transcription start site for the piu operon by reverse transcriptase mapping. 5'-32P-labeled DNA oligomer 13R, complementary to the noncoding strand within the piuB ORF, served as the primer. The results of the primer extension (Fig. 7) indicate that transcription in the piuBCDA promoter region is initiated at coordinate –42 (G) relative to the ATG of the piuB ORF. The ability of RitR to shift the mobility of the piuB promoter region was tested by using a 5'-32P-labeled 217-bp DNA fragment upstream of piuB between coordinates –107 and 110 relative to the transcriptional start site of piuBCDA. The labeled DNA was incubated with GST-RitR, and the resultant complexes were analyzed by PAGE. A GST-RitR concentration of 60 nM was enough to show evidence of retardation (Fig. 7B). Preincubation of GST-RitR with 50 mM acetyl phosphate or phosphoramidate did not alter the affinity of GST-RitR for the piu promoter (data not shown). These data suggested that GST-RitR bound specifically to the piu promoter.
|
|
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
ritR) indicated that one role of RitR is to repress transcription of these genes. The finding that piuC and piuD were not differentially expressed could reflect selective degradation of the piuBCDA message at internal sites in the transcript. Interestingly, a similar effect was attributed to differential instability of the four-gene transcript in the Vibrio anguillarum fat Fe transporter operon, which the piu transporter closely resembles (10, 57). In addition, further experiments showed that RitR binds physically at three sites upstream of piuB to a specific AT-rich sequence. Collectively, our data suggest a role in iron regulation for RitR in S. pneumoniae.
Uptake of iron in the form of hemin.
Tai et al. (50) reported that PiuB binds hemin with a high affinity in vitro, suggesting that this is the form in which PiuBCDA acquires and possibly transports iron. Indeed, compared with R800, RU402 (ritR) exhibited 5 orders of magnitude difference for Fe2+ and 3 orders of magnitude difference for hemin (but not hemoglobin or Fe3+) with respect to streptonigrin susceptibility. Interestingly, a heme oxygenase determinant does not seem to be present in the pneumococcal genome (52). Given this observation, we can only assume that iron somehow enters the cell in the free ionic form until sufficient data become available to pinpoint the precise mechanism(s) of iron transport. In addition, we cannot rule out the possibility that the repression of iron storage protein genes (e.g., dpr [log2 ratios of 2.5 and 6.2 in the R800-versus-RU402 and tetracycline-induced arrays, respectively] [Table 4]) could affect intracellular iron concentrations and thus streptonigrin susceptibility. Isotopic Fe2+ uptake studies are needed to resolve these issues.
The existence of two additional iron transporter gene clusters, namely, pit and pia, may provide other options for iron acquisition and for interpreting the streptonigrin effect shown in Fig. 6A and B (5, 6). Interestingly, a RitR binding site has also been identified upstream of the pia gene cluster between its predicted –35 and –10 promoter elements, suggesting an additional iron-regulatory role for RitR (data not shown). The additive effect of piaA and piuA in pathogenesis is consistent with the observations of Brown et al. (8), who noted that mice immunized with recombinant PiuA and/or PiaA were protected against a systemic challenge.
Regulation of RitR. The present studies suggest that RitR represses the piu iron transporter, but the precise mechanism by which RitR itself is regulated is unknown. Four possible mechanisms of regulation are (i) control of the activation of RitR by phosphorylation, (ii) control of the amount of RitR that is synthesized, (iii) interaction of RitR with another protein (e.g., phosphatase), and (iv) modulation of RitR activity by noncovalent ligand-protein interaction. Neither a cognate kinase for RitR nor a mechanism by which S. pneumoniae senses iron has been found. Additionally, at this time it is not clear whether a cognate kinase even exists or if phosphorylation plays any role in the regulatory activity of RitR. Indeed, we were unable to alter the affinity of RitR for the piu promoter when we attempted to phosphorylate it by addition of small phosphate donors (Fig. 7). Even though RitR appears to contain a conserved aspartate residue, these data bring into question whether this response regulator has evolved another mechanism of regulation (e.g., growth rate-dependent regulation); however, a role for phosphorylation of RitR cannot be ruled out at this time. Interestingly, C. diphtheriae has a TCST system, ChrAS, that has been reported to sense heme through the histidine kinase ChrS and to initiate transcription of hmnO (encoding heme oxygenase) following activation of the response regulator ChrA (45). It is tempting to speculate that the orphan RitR works in a fashion similar to ChrA through 1 of the 13 histidine kinases encoded in the S. pneumoniae genome, sensing iron-hemin in the host and signaling for derepression of the piu transporter.
Coregulation of gnd-ritR and H2O2 toxicity. gnd and ritR are cotranscribed as a single 2-kb mRNA (Fig. 4). Given this observation, elevated expression of gnd-ritR would be expected to repress iron uptake while stimulating expression of the pentose phosphate pathway, which is responsible for (i) generating NADPH, which provides reducing power for detoxification of reactive oxygen intermediates, and (ii) synthesizing nucleotides needed for repair of Fe-catalyzed damage of DNA (46, 59).
S. pneumoniae is able to produce high levels of H2O2, which can reach millimolar concentrations. This effect has been shown to aid the pneumococcus in colonization of the nasopharynx by inhibiting the growth of competitors, such as Haemophilus influenzae (36). A more direct use of H2O2 as a weapon by S. pneumoniae is suggested by its cytotoxic effect on epithelial cells (14). The observations reported in our studies are relevant to the findings reported by Pericone et al. (37), who suggested that catalase-deficient S. pneumoniae is able to survive toxic concentrations of H2O2 in the presence of intracellular iron concentrations exceeding the concentration in E. coli. Possible contributing factors that were suggested included (i) an enhanced ability to repair DNA damaged by ROIs and (ii) the presence of an H2O2 resistance protein, such as Dpr, that sequesters iron. Both of these mechanisms are consistent with our experimental findings.
The enhanced ability to repair DNA damaged by ROIs could be mediated by gnd-ritR activation. The pentose phosphate could provide the cell with NADPH for detoxification of Fe-catalyzed synthesis of ROIs and synthesis of nucleotides for excision repair, while RitR could contribute to a reduction in Fe-catalyzed ROI synthesis by reducing iron uptake during exponential growth under high-iron conditions. In support of this interpretation, the resistance of Deinococcus radiodurans to UV light and H2O2 was shown by Zhang et al. to depend on an intact pentose phosphate pathway (59). In a comparison of a zwf+ (Zwischenferment, glucose-6-phosphate dehydrogenase) strain and a zwf strain of D. radiodurans, Zhang et al. noted 1-order-of-magnitude reduced survival of the zwf mutant in 0.03% (10 mM) H2O2. In contrast, only a minimal difference was seen in a comparison between an fda+ (aldolase) strain and a fda strain in which the Embden-Meyerhoff pathway was interrupted. Moreover, our array data indicate that mutY RNA levels are activated in the presence of RitR. The S. pneumoniae mutY gene encodes an A/G glycosylase responsible for removal of 7,8-dihydro-8-oxyoguanine (44), a major product of oxidative damage to DNA (18, 29). Collectively, these data support our conjecture that S. pneumoniae could possess a greater ability to repair DNA than E. coli possesses that enables it to survive in the presence of high concentrations of H2O2. Experiments to test whether the pentose phosphate pathway plays a role in DNA repair in S. pneumoniae similar to the role that it plays in D. radiodurans are currently in progress.
The second H2O2 resistance factor suggested by Pericone et al. (37) is the protection of DNA by Dpr, a protein that sequesters intracellular Fe2+ and prevents H2O2 toxicity. Indeed, our array data show that there is a marked increase during exponential growth in the mRNA levels of the putative H2O2 resistance orthologs adhE, encoding an iron-binding alcohol dehydrogenase, and dpr. This observation indicates a possible role for these genes in S. pneumoniae so that it can avoid DNA damage caused by interactions with Fe-catalyzed ROIs. Interestingly, neither Pericone et al. (37) nor workers in our lab (data not shown) could obtain a dpr deletion mutant, suggesting that the Dpr protein is required for viability of S. pneumoniae. Concomitant with repression of iron transporter synthesis by RitR, elevated dpr expression would be consistent with a model in which Dpr sequesters iron that enters the cell despite repression of iron transporter synthesis. However, it is not clear at this time whether RitR directly regulates dpr or adhE or whether their activation is an indirect effect of excess iron entering the cell in RU402 through another signaling system.
Sugar transport-ugd regulation by ritR. Three components belonging to three separate sugar transport systems were observed to be strongly up-regulated in the presence of RitR during log-phase growth. These components included two sugar-binding proteins and a permease belonging to a putative ribose/galactose sugar transporter (Table 4). Interestingly, D-ribose can be metabolized only via the pentose phosphate pathway (46), which is directly associated with ritR because this gene is cotranscribed with gnd. Additionally, ugd was strongly down-regulated in the presence of ritR. In Salmonella enterica serovar Typhimurium, Ugd is regulated directly by three different TCST systems: PmrA/B, which senses Fe3+; PhoP/Q, which senses low concentrations of Mg2+; and RcsC/B/YojN, which senses a variety of signals, including osmolarity and temperature (32). Ugd catalyzes the conversion of UDP-D-glucose to UDP-D-glucuronic acid, a necessary step in the synthesis of capsular polysaccharides in S. pneumoniae type 3 (3), and this process has been shown to be negatively regulated by tyrosine kinase (CpsD)-mediated phosphorylation (30). A possible interpretation of these observed regulatory changes is that while up-regulation of RitR can potentially activate the pentose phosphate pathway through gnd, it can also regulate sugar uptake and provide the pentose phosphate pathway with substrates while repressing alternative pathways for sugar utilization (e.g., capsule biosynthesis through ugd). Interestingly, a RitR consensus binding sequence is present in the promoter region of ugd located 151 bp from upstream of its start codon, suggesting a direct role for RitR regulation of this gene.
Compared to other organisms, little is known about iron homeostasis and oxidative stress regulation in S. pneumoniae. In this report we show that RitR represses the piu iron uptake operon and is linked to the expression of genes involved in the oxidative stress response. In addition, we show that ritR is cotranscribed with gnd, which encodes a critical enzyme of the pentose phosphate pathway that produces nucleotides and reducing power for the cell. In the absence of an identified cognate kinase, it will be of interest to learn what factors regulate RitR. At this time we can only point to the linkage of ritR to gnd.
|
ACKNOWLEDGMENTS |
|---|
Tyson Park provided expert technical assistance. For numerous helpful discussions we thank the staff of the University of Wisconsin Gene Expression Center, including Sandra Splinter Bon-Durant, John Luecke, and Nelson Wayne Davis. We also thank Donald Morrison for helpful discussions and for plasmid pEVP3 and M. Stieger for plasmids pRK01 and pRK02. We thank Christina Kendziorski for the use of her program EBarrays to identify differentially expressed genes.
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
