当前位置: 首页 > 医学版 > 期刊论文 > 基础医学 > 感染与免疫杂志 > 2006年 > 第2期 > 正文
编号:11255363
Expression Microarray and Mouse Virulence Analysis of Four Conserved Two-Component Gene Regulatory Systems in Group A Streptococcus
     Center for Human Bacterial Pathogenesis Research, Department of Pathology, Baylor College of Medicine, Houston, Texas 77030

    Center for Molecular and Translational Human Infectious Diseases Research, The Methodist Hospital Research Institute, and Department of Pathology, The Methodist Hospital, Houston, Texas 77030

    ABSTRACT

    Group A streptococcus (GAS) is a gram-positive human bacterial pathogen that causes diseases ranging from relatively mild epithelial cell surface infections to life-threatening invasive episodes. Much is known about the extracellular molecules that contribute to host-pathogen interactions, but in contrast, far less information is available about regulatory genes that control the expression of individual or multiple GAS virulence factors. The eight GAS genomes that have been sequenced have 12 conserved two-component gene regulatory systems (TCSs), but only 3 of these 12 have been studied in detail. Using an allelic replacement strategy with a nonpolar cassette, we inactivated the response regulator of four TCSs that have only weak homology with TCS genes of known or inferred function in other bacteria. The mutant strains were analyzed by expression microarray analysis at four time points and tested in two mouse infection models. Each TCS influenced expression (directly or indirectly) of 12 to 41% of all chromosomal genes, as assessed by growth in Todd-Hewitt broth and a custom Affymetrix GeneChip. None of the isogenic mutant strains was significantly altered for mouse virulence based on intraperitoneal inoculation. Similarly, compared to the wild-type strain, there was no significant difference in skin lesion size for three of the four mutants. In contrast, the M5005_Spy_0680 mutant strain produced significantly larger abscesses after subcutaneous inoculation into mice, consistent with a hypervirulence phenotype. The mutant strain had significantly higher in vitro expression of several proven and putative virulence genes, including scpA, encoding a peptidase that inactivates complement protein C5a. Together, the data provide new information about previously uncharacterized GAS TCSs.

    INTRODUCTION

    Group A Streptococcus (GAS) is a gram-positive bacterium that causes a wide array of human infections at different anatomic sites. The severity of these infections varies from mild skin or mucous membrane involvement to invasive life-threatening episodes characterized by invasion of deep tissues. In the last 60 years, considerable effort has been devoted to studying GAS extracellular molecules that contribute to host-pathogen interactions. Many important virulence factors that promote disease, including cytotoxins, pyrogenic toxin superantigens, proteases, and molecules that affect innate and acquired immune functions, have been identified. In contrast, far less information is available about regulatory genes that control the expression of individual or multiple GAS virulence factors. For example, in the eight published GAS genomes (2, 3, 6, 9, 24, 32, 33, 36), approximately 125 genes are annotated as involved in signal transduction or information processing. However, the great majority of these genes have not been studied, which means that much remains to be learned about regulatory genes and control of gene expression in GAS.

    To sense and respond to changes in external conditions, bacteria often use a specialized gene pair referred to as a two-component gene regulatory system (TCS). Changes in the environment are detected by a membrane-localized sensor kinase that transduces a signal to a receiver protein (response regulator), a process resulting in altered gene expression. TCSs have been well studied in many pathogenic bacteria, and they are known to regulate expression of many key virulence factors, such as extracellular toxins.

    The eight GAS genomes that have been sequenced have 13 TCSs, including 12 that are well conserved in all strains (2, 3, 6, 9, 24, 32, 33, 36) (Fig. 1A). Only 3 of these 12 TCSs (fasBCA, covRS, and ihk/irr) have been studied in detail, and information is available about their involvement in host-pathogen interactions (7, 8, 15, 40). Genetic analysis of fasBCA mutant strains revealed that this TCS assists in tissue invasion and destruction by regulating expression of genes encoding secreted virulence factors, including streptokinase, streptolysin S fibrin, and fibrinogen binding proteins. Expression of the fasBCA genes in exponential phase is affected by environmental signals, such as a temperature change from 29°C to 37°C (34). The CovR/S TCS has been studied extensively and is known to be a major negative regulator of virulence genes (5, 7, 16). For example, inactivation of the CovR/S TCS significantly increases virulence for mice and resistance to complement-mediated opsonophagocytic killing in vitro (7, 16) and accordingly increases expression of key virulence factors, such as hasABC, skaA, sagA, speB, and speMF (5, 7). Expression microarray analysis found that CovR/S directly or indirectly influences expression of 15% of genes in the GAS genome (7). Expression of this TCS is altered by the external concentration of Mg2+ ions (11). It has been suggested that a low concentration of Mg2+ in extracellular body fluids prevents derepression of CovR-regulated genes and allows for their full expression in host tissues (11).

    The Ihk/Irr TCS recently has been studied in detail, motivated by the observation that its expression was up-regulated significantly during interaction of GAS with human polymorphonuclear leukocytes (41). Inactivation of irr detrimentally affected virulence in mouse models of soft tissue infection and bacteremia. Microarray analysis revealed that ihk/irr influences expression of many genes involved in maintenance of cell redox status and cell wall biosynthesis. Consistent with this, an irr mutant strain was rapidly killed by antibacterial peptides that target the cell envelope, suggesting that cell wall metabolism plays a crucial role in GAS survival in vivo (40).

    In the present work, we used isogenic strains, expression microarray analysis, and mouse infection models to investigate four TCSs of unknown function in GAS. We focused our studies on TCSs that have only weak homology with TCS genes of known or inferred function in other bacteria.

    MATERIALS AND METHODS

    Bacterial strains, routine growth, and oligonucleotide primers. Serotype M1 strain M5005 was used to construct all mutant derivative strains. Strain M5005 is genetically representative of serotype M1 strains responsible for the great majority of contemporary episodes of disease caused by GAS of this serotype (12, 13, 23). The strain has been very well characterized and used extensively in previous studies on GAS pathogenesis (7, 8, 35, 39). Recently, the genome of strain M5005 has been sequenced (36).

    Strain M5005 and its isogenic derivatives were grown in Todd-Hewitt broth (Difco Laboratories) supplemented with 0.2% yeast extract (THY medium) at 37°C in an atmosphere of 5% CO2-20% O2. THY agar or tryptose agar with 5% sheep blood (Becton Dickinson) was used as solid medium. THY agar supplemented with spectinomycin (150 μg/ml) was used for selection of antibiotic-resistant mutant derivative strains. Cloning experiments were performed with Escherichia coli Nova Blue (Novagen). Ampicillin (100 μg/ml) or spectinomycin (150 μg/ml) was used for selection of E. coli clones when required. The oligonucleotide primers used to construct mutant strains are presented in Table 1.

    DNA techniques. Restriction and modification enzymes were purchased from New England BioLabs or Fermentas. Plasmid DNA from E. coli was isolated with a Nucleo Spin Plasmid kit (Macherey Nagel). Chromosomal DNA was isolated from GAS using a Tissue Nucleo Spin kit (Macherey Nagel) as described by the manufacturer, with slight modification. Before lysis, GAS strains were frozen at –70°C, thawed, and incubated with mutanolysin (500 U/ml) and lysozyme (1 mg/ml) for 30 min at 37°C. Southern hybridization was performed with an ECL system according to the manufacturer's instructions (Amersham).

    Transformation of GAS. Strain M5005 was grown overnight at 37°C in THY broth supplemented with 40 mM L-threonine and 250 mM sucrose, subcultured to fresh medium, and grown to an optical density at 600 nm (OD600) of 0.2 to 0.25. Bacteria were collected by centrifugation, washed three times with ice-cold 0.5 M sucrose, and resuspended in 0.5 M sucrose with 20% glycerol. Competent cells (100 μl) were mixed with dialyzed DNA and transferred to a cooled cuvette with a 0.1-cm electrode gap. After 30 min of incubation, the bacteria-DNA mixture was subjected to a pulse of 1.8 kV, 25 μF, 400 . The bacteria were mixed immediately with 1 ml THY containing 250 mM sucrose, incubated for 2 to 3 h at 37°C, and plated onto THY medium containing spectinomycin.

    Construction of isogenic mutant strains with inactivated TCS response regulator genes. Four isogenic mutant strains were constructed, each having an inactivated response regulator gene (M5005_Spy_0680, M5005_Spy_0785, M5005_Spy_0830, and M5005_Spy_1281) of a TCS (Fig. 1B). These genes are designated SPy0874, SPy1062, SPy1106, and SPy1556, respectively, in the genome of strain SF370 (Fig. 1A). The gene numbers differ between strains SF370 and M5005 because the genome of the latter strain contains several large blocs of prophage DNA not present in the genome of strain SF370 (6, 36). To construct each mutant strain, the coding sequence of the response regulator gene was replaced with a spectinomycin resistance cassette by a double crossover strategy. DNA fragments encoding the upstream region of four response regulator genes were amplified by PCR with primer pairs (Table 1) 0680LF-0680LR (530 bp), 0785LF-0785LR (557 bp), 0830LF-0830LR (562 bp), and 1281LF-1281LR (465 bp). The purified gene fragments were digested with NdeI/XmaI and cloned into pUC19 (Fermentas) digested with appropriate enzymes to yield plasmids p0680-5', p0785-5', p0830-5', and p1281-5'. The downstream region of each gene encoding the target response regulators was amplified with primers 0680RF-0680RR (490 bp), 0785RF-0785RR (436 bp), 0830RF-0830RR (469 bp), and 1281RF-1281RR (466 bp), cloned into pSTblue-1 vector (Novagen), and digested with HincII/AvrII to generate plasmids p0680-3', p0785-3', p0830-3', and p1281-3'. The spectinomycin resistance cassette (spc) containing the add9 gene, which was shown to have no polar effects on downstream genes (19), was excised from plasmid pSL60-2 with SmaI and cloned into the PmlI site of each plasmid containing the downstream fragments of the response regulator genes. The orientation of the spc cassette was determined by PCR, and plasmid clones containing the spc cassette and downstream insert present in the same orientation were used for subsequent cloning steps. Fragments with the spc cassette fused with GAS DNA fragments were excised with Ecl136II/SnaBI and subcloned into the SmaI site of plasmids p0680-5', p0785-5', p0830-5', and p1281-5' containing the upstream region of each gene to be inactivated. The resulting plasmids (pM0680, pM0785, pM0830, and pM1281) have the upstream and downstream regions flanking each response regulator with the spc cassette cloned between them rather than the gene encoding the response regulator. The final recombinant DNA constructs were linearized with ScaI and transformed into electrocompetent strain M5005. GAS transformants were selected on solid medium containing spectinomycin. Chromosomal DNA isolated from spectinomycin-resistant colonies was used in PCR screening with pairs of primers 0680L-0680R, 0785L-0785R, 0830L-0830R, and 1281L-1281R. Primer pairs were designed to complement the chromosomal DNA sequence located outside of the targeted integration site. Putative recombinants were confirmed to contain the correct mutation by PCR analysis, DNA sequencing, and Southern hybridization using the spectinomycin cassette as a probe (Fig. 2). Successful inactivation of the target gene also was confirmed by the results of the microarray experiment, which showed the absence of transcripts of the inactivated response regulator gene (see below).

    Expression microarray analysis. Expression microarray analysis was performed using a custom-made Affymetrix chip that has been described in detail (8, 39). RNA samples used for microarray analysis were prepared as described previously (8, 39). Briefly, bacteria were grown in groups of five cultures (four mutant strains and the isogenic wild-type parental strain M5005) in fivefold redundancy (total number of cultures = 25) using the same lot of medium. Hence, for each strain, five independent GAS cultures (five biological replicates) served as the source of RNA used for array analysis. Bacteria were harvested at four time points: when the OD600 was 0.2 (early logarithmic phase), when the OD600 was 0.5 (mid-logarithmic phase), when the OD600 was 1.1 to 1.2 (late logarithmic phase), and when the OD600 was 1.8 to 1.9, a time 3 h after the bacteria entered stationary phase. The total number of RNA samples analyzed was 100 (five strains times five biologic replicates times four time points). The bacteria were mixed with RNA Protect reagent (QIAGEN), and GAS cells were collected by centrifugation and stored at –80°C until processing. Labeled cDNA (targets) used for hybridization with the custom chip were prepared according to the procedure described by Affymetrix for use with Pseudomonas aeruginosa (1) with modifications described recently (8, 39).

    To minimize nonbiological variability caused by the multistep sample preparation and hybridization procedure, the targets were prepared according to a complicated experimental protocol designed by Partek Inc. and used in other studies (8, 39). All bacterial samples prepared on one day (5 strains x 4 time points = 20) were processed at the same time using 96-well plates for all steps of the protocol, from RNA isolation to labeling. All samples were placed near the center of the 96-well plate to minimize possible spurious variation induced by a plate-edge effect. Similarly, to avoid potential "right-to-left," "top-to-bottom," or "center-to-edge" spurious effects, the samples were arranged in a computer-generated matrix that took into account experimental parameters (strain and growth phase). This strategy minimizes the possibility that trends identified for changes in gene transcripts are confounded by unrelated parameters, such as culture batch or steps in target preparation. Each group of 20 samples prepared simultaneously was hybridized simultaneously with the custom GeneChips. Other conditions or parameters that in principle can detrimentally affect the quality of the resulting data (chip lot, wash batch, position, or scan order) were randomly assigned.

    Chip hybridization data were acquired and normalized using Affymetrix GeneChip Operating Software (GCOS). Hybridization intensity values were normalized to the mean intensity of all GAS genes present on the chip using GCOS version 1.0, and only genes with a "present" signal were included for further analysis. Tables generated by GCOS containing normalized hybridization values obtained from all 100 chips were exported to Partek Pro 6.0 package (Partek Inc.) for statistical validation of data and subsequent analysis. Hybridization data were analyzed using a three-way analysis of variance (ANOVA) and linear contrasts to identify genes differentially expressed between each knockout strain and wild-type strain M5005. The blocking factors used for the ANOVA were strain, growth phase, and culture batch. Using this approach, factors such as wash batch, chip lot, position on plate, etc., can become covariate in the ANOVA to reduce noise. A PartekPro principal component analysis (PCA) and visualization system was used to assess chip quality and chip-to-chip variability. Input information combined hybridization intensity values and information about sample preparation and hybridization.

    All expression microarray data have been deposited in the Gene Expression Omnibus (GEO) database at NCBI at http://www.ncbi.nlm.nih.gov/geo/.

    TaqMan confirmation of microarray analysis. To confirm that obtained microarray data represent true changes in gene expression in mutant strains, we validated the data with quantitative real-time PCR (TaqMan). From transcripts detected by microarray analysis as differentially expressed, we selected at random several (8 to 11 transcripts per mutant strain) of them for analysis. cDNA for TaqMan confirmation was generated from the RNA samples used for microarray analysis (two of five biological replicates) using the same procedure as for microarrays. TaqMan reactions for each gene tested were prepared in quadruplicate for each of two biological replicates and were amplified using Platinum Quantitative PCR SuperMix-UDG (Invitrogen). The transcript amounts in each mutant were standardized to an internal control gene (proS) and compared with standardized expression in the wild-type strain (CT method).

    Mouse infection experiments. The two mouse models used for infection experiments have been described extensively in previous studies (18, 20, 21). GAS strains used for mouse infection studies were grown in THY medium to exponential phase (OD600 0.5), harvested, washed twice with ice-cold phosphate-buffered saline, and frozen at –70°C in aliquots. Thawed aliquots of strain M5005 and the isogenic mutant strains were adjusted to the same CFU/ml by diluting with phosphate-buffered saline. The number of CFU used for mouse inoculation was determined by plating on sheep blood agar. Male immunocompetent hairless mice (4 to 6 weeks old) of strain Crl:SKH1-hrBR (Charles River) were used for subcutaneous inoculation studies. The animals were anesthetized with isoflurane and inoculated with 2 x 107 CFU of GAS in 100 μl. To monitor abscess formation and infection progression, animal weight and abscess size were measured every day for the first week and twice a week thereafter for a total of 28 days. Abscess length (L) and width (W) values were used to calculate abscess area (A = [L/2] x [W/2]) and volume (V = [4/3][L/2]2 x [W/2]) using the equation for a spherical ellipsoid (20). Outbred CD-1 Swiss male mice (4 to 6 weeks old) (Harlan) were used for intraperitoneal inoculation with a dose of 5 x 107 CFU in 250 μl. Near-mortality was monitored every 2 h for the first 48 h after infection and then every 6 h for the next 5 days.

    RESULTS

    Construction of isogenic mutant strains impaired in TCS expression. As a first step toward probing the genes regulated (directly or indirectly) by four GAS TCSs of unknown function, we inactivated the response regulator of each TCS. All four isogenic mutant strains were derived from the parental M1 serotype strain M5005, whose genome has been recently sequenced (36). We used an allelic replacement strategy, in which the genes were inactivated using a spectinomycin resistance cassette, which was shown previously to have no polar effects on downstream genes (19). PCR analysis, DNA sequencing, and Southern hybridization confirmed that the proper genetic constructs had been made (Fig. 2).

    There was no observable difference in colony phenotype or in vitro grow rate between the wild-type strain and the four mutant strains on blood agar or tryptose agar (data not shown). For example, there was no difference in colony size, zone of hemolysis, or mucoid appearance of the colonies. Similarly, there was no difference between the wild-type and mutant strains in bacterial chain length (data not shown). Consistent with these observations, the growth rate in THY medium of each mutant strain was identical to the wild-type parental strain (Fig. 3).

    Expression microarray analysis of the mutant strains. TCSs in pathogenic bacteria, including GAS, are well known to influence expression (directly or indirectly) of very large numbers of genes. To identify the spectrum of genes under the influence of each TCS, we analyzed gene transcripts of the wild-type parental strain and each mutant strain at four time points (representing early, mid-, and late logarithmic and stationary growth phases) of in vitro growth using a custom GeneChip manufactured by Affymetrix Inc. (8, 17, 39). The experimental design and procedures used minimized factors that can detrimentally affect data quality and reproducibility. To provide enough data for robust statistical analysis, we used five sample replicates for each strain at each of the four growth phase time points (see Materials and Methods). Thus, data from 100 chips were used for the expression array analysis.

    Chip quality and chip-to-chip variability were assessed using a PartekPro PCA and visualization system. Importantly, the PCA discriminated very well between the data from the chips representing early logarithmic, mid-logarithmic, late logarithmic, and early stationary growth phases (Fig. 4). These results indicated that the transcriptome profile data are of high quality and demonstrated extensive remodeling of the transcript profile over time during in vitro culture.

    The microarray data were confirmed using quantitative PCR. cDNA generated from two biological replicates used for microarray analysis served as a template in TaqMan reactions. Figure 5 presents results of correlation analysis between expression data obtained for 35 genes in microarray and TaqMan experiments. For all four studied mutants, both methods delivered highly similar results, with R2 values after correlation analysis between 0.8712 and 0.9814.

    Overview of the expression array data. Analysis of the transcriptomes of all five strains (four mutant strains and the wild-type parental strain) revealed that 1,632 transcripts were detected by hybridization with the chip. For each of three TCS mutant strains (M5005_Spy_0680, M5005_Spy_0785, and M5005_Spy_0830), expression of about 15% of all genes was significantly altered compared to the wild-type strain at one or more time points. In contrast, the transcriptome of mutant strain M5005_Spy_1281 was far more altered compared to the wild-type strain, with 41% of transcripts detected after hybridization affected at one or more time points (Fig. 6). Changes in transcript levels correlated with growth phase in three of the four mutant strains (Fig. 7, left panel for each category; see Fig. S1 in the supplemental material). For example, in the M5005_Spy_0680 mutant strain, the majority of affected transcripts occurred in the late logarithmic and stationary phases. Similarly, a massive change in transcript profile was observed in the M5005_Spy_1281 mutant strain in the stationary phase, with over 70% of all changes occurring at that time point. Transcripts in the M5005_Spy_0785 mutant strain were mostly affected in the early exponential and stationary phases. Finally, mutant strain M5005_Spy_0830 differed prominently from the other three mutant strains, in that no correlation was observed between growth phase and altered transcript profile (Fig. 7).

    Analysis of transcripts in the M5005_Spy_0680 strain revealed that more genes were down-regulated (n = 206) than up-regulated (n = 136) at one or more of the time points analyzed. The transcript representing the M5005_Spy_0681 kinase was up-regulated in the mutant strain, suggesting that its cognate response regulator acts as a repressor of its own operon (Fig. 1B).

    Similar to mutant strain M5005_Spy_0680, more genes were down-regulated (n = 181) than up-regulated (n = 106) in the M5005_Spy_0785 mutant strain. However, the transcript encoding the cognate sensor kinase was down-regulated, suggesting that it acts as an activator of its own operon (Fig. 1B). In addition, the transcripts representing an apparent operon (M5005_Spy_0780-0783) located directly upstream of this TCS were greatly down-regulated (200-fold), suggesting that M5005_Spy_0785 also positively regulates expression of this operon encoding carbohydrate metabolism proteins (Fig. 1B).

    Inactivation of the M5005_Spy_0830 response regulator affected the smallest number of GAS transcripts, with no bias toward down-regulated (n = 115) or up-regulated (n = 103) genes. In general, similar numbers of transcript changes occurred in all growth phases. The level of down-regulation was quite high in the strain with mutations of two contiguous genes (M5005_Spy_0832-0833) in an operon encoding a putative NAD-dependent malic enzyme (40-fold decrease) and malate-sodium symport (27-fold decrease). The transcript level of M5005_Spy_0831, encoding the sensor kinase component of this system, was decreased in the mutant strain, consistent with the idea that the response regulator is an activator of this TCS (Fig. 1B).

    Transcript factor changes for genes in the M5005_Spy_1281 strain varied from –8.1-fold (M5005_Spy_0149/SPy0175, encoding component of the phosphotransferase system [PTS]) to +5.4-fold for araD/SPy0179. The studied response regulator presumably acts as repressor of its own transcription, as the transcript encoding the sensor kinase was up-regulated (Fig. 1B).

    Gene categories affected in mutant strains. The annotated genes in the genome of strain M5005 were assigned to 16 functional categories. The percentage of annotated genes in each functional category was compared with the percentage of transcripts affected during at least one time point in the four analyzed mutant strains, regardless of growth phase (Fig. 7). This type of comparison can be useful because it can suggest a role for regulators of unknown function on the basis of over- or underrepresented gene categories. In general, for most gene categories, the percentage of affected genes in the mutant strains did not differ dramatically from the percentage of annotated genes in the wild-type strain. However, some differences were observed (Fig. 7, yellow panels in each functional category). For example, a prominent feature of the transcriptome of strain M5005_Spy_0680 was the percentage of affected genes involved in carbohydrate and nucleotide metabolism. The number of affected transcripts was especially high in the transition from logarithmic to stationary phase (Fig. 7, left panels).

    Mouse virulence of the four TCS mutant strains. Inactivation of many TCSs in pathogenic bacteria alters virulence due to changes in gene expression, including but not limited to genes encoding virulence factors such as extracellular or cell wall-anchored proteins (5, 7, 15, 41). Inactivation of bacterial TCSs usually decreases virulence, but exceptions to this general observation have been reported (7, 28). Two well-described mouse models of infection were used to test the hypothesis that inactivation of these four GAS TCSs alters virulence. We first assessed the mouse virulence of the mutant strains by intraperitoneal inoculation. No significant virulence difference was observed between the wild-type parental strain and the four isogenic mutant strains as assessed by Kaplan-Meier survival curve analysis (data not shown). Next, the virulence of the four mutant strains relative to the wild-type strain was analyzed in a soft-tissue model of GAS infection. These experiments were conducted because it is possible that genes required for maximum virulence in this model differ from the intraperitoneal injection model. No significant difference in virulence was observed for three of the four mutant strains (M5005_Spy_0785, M5005_Spy_0830, and M5005_Spy_1281). In contrast, mutant strain M5005_Spy_0680 caused significantly larger abscesses than wild-type parental strain M5005, consistent with a hypervirulence phenotype in this model (Fig. 8).

    Altered transcriptome of the M5005_Spy_0680 hypervirulent mutant strain. To gain initial insight into the possible molecular basis of the large-abscess phenotype of strain M5005_Spy_0680, we compared the transcripts affected in this mutant strain with the wild-type parental organism, with special focus on genes potentially involved in virulence. The transcript level of several genes known to enhance virulence was significantly increased in the mutant strain, including scpA, sagGI, silD, and sic. In addition, the gene (prtS) encoding a recently described secreted protease that cleaves and inactivates interleukin-8 (4) was up-regulated in the early logarithmic phase of growth. The largest category of genes with altered transcripts was genes of unknown function, with the majority of changes observed in the stationary and/or late logarithmic growth phases. Although it is not known whether any of these genes encode virulence factors, we note that four of them (M5005_Spy_0354, M5005_Spy_0355, M5005_Spy_0666, and M5005_Spy_0721) have typical or probable gram-positive secretion signal sequences. Importantly, these four transcripts were all increased, consistent with an enhanced-virulence phenotype.

    As noted above, inactivation of the M5005_Spy_0680 response regulator also affected the expression of multiple genes apparently involved in bacterial metabolism. For example, decreased expression of M5005_Spy_0025-0031 genes (purMNDEK) and M5005_Spy_0347-0349 genes (nrdFIE), involved in nucleotide metabolism, and genes involved in arginine utilization (M5005_Spy_1237-1238/SP1506-1507) were observed. In addition, carbohydrate utilization genes were prominently affected, including several PTS transport systems and genes apparently involved in maltose (M5005_Spy_1055-1067/SP1291-1304) and tagatose (M5005_Spy_1636/SP1921) utilization pathways. These and other findings are summarized in Fig. 9.

    DISCUSSION

    Our primary goal in this study was to provide new information about four GAS TCSs of unknown function. Prior to this study, only 3 of the 12 conserved GAS TCSs had been studied for a potential role in host-pathogen interactions. FasBCA influences expression of adhesins, streptokinase, and streptolysin S (15). Ihk/Irr is involved in resistance of GAS to phagocytosis and killing by human PMNs (40). CovR/CovS is the best-studied GAS TCS, and it is known to regulate expression (directly or indirectly) of many key virulence genes and multiple aspects of GAS physiology (5, 7, 8, 16). We focused our efforts on four TCSs of unknown function that are conserved in all eight sequenced GAS genomes (2, 3, 6, 9, 24, 32, 33, 36). Each mutant strain had no observable phenotype when grown in vitro in rich liquid medium or sheep blood agar. Specifically, there was no change in colony morphology or growth rate during in vitro growth in THY medium. These findings are consistent with previous studies showing that inactivation of TCS from GAS and other bacteria does not necessarily alter phenotype such as growth rate in rich liquid medium (5, 14, 31).

    Analysis of the in vitro growth phenotype of mutant strains can yield useful data, but studies of this type do not address the role of specific genes in pathogen interaction with an intact host. Thus, we used two well-characterized mouse models to test the hypothesis that inactivation of each of these four TCS genes altered virulence. The M5005_Spy_0680 mutant strain caused significantly larger skin lesions in mice than the wild-type parental strain M5005, whereas the other mutants were not altered in mouse virulence. Although we did not address the molecular basis of increased virulence of the M5005_Spy_0680 mutant strain, we identified clues that may help to explain this phenotype. Expression of 310 genes was directly or indirectly influenced by this TCS during at least one of the time points analyzed, including several (scpA, sagGI, silD, and sic) known to participate in GAS pathogenesis. In addition, we note that the prtS gene encoding a protease that inactivates interleukin-8 (4) was up-regulated in the early logarithmic phase of growth, a process that could contribute to immune evasion in vivo. Expression also was increased for 24 hypothetical proteins with no homologs in other bacteria. None of these 24 up-regulated genes encoded proteins with an LPXTG cell wall-anchoring motif, a common feature of many virulence factors made by gram-positive bacteria. However, 5 of these 24 proteins have putative membrane-spanning motifs, raising the possibility that they are involved in transport of molecules (including virulence factors). In addition, four up-regulated proteins have gram-positive secretion signal sequences and thus may be located extracellularly. We speculate that study of the involvement of these proteins in host-pathogen interaction might be fruitful.

    Global analysis of in vitro transcription. Inactivation of each of the four response regulators affected expression of a very large number of genes, consistent with previous transcriptome studies of CovR/CovS and Ihk/Irr (7, 40). For example, the percentage of affected genes during at least one experimental time point varied from 12% to 41% of genes, as observed for the M5005_Spy_1281 mutant strain. In addition to the large number of affected genes, changes in transcription profiles were clearly growth phase dependent. Most genes were affected at only one or two of the time points analyzed, suggesting that each TCS acts primarily at a distinct phase of growth. Virtually all transcription changes were relatively modest, rarely exceeding a twofold difference (see Fig. S1 in the supplemental material). Taken together, these two observations suggest that most of the changes in gene expression were indirect effects mediated by secondary effectors rather than primary effects caused by direct interaction of the response regulator with many different promoter regions. It is also possible that some of the changes in transcript level were due to cross talk (a sensor kinase phosphorylates a noncognate response regulator) or cross-regulation (a sensor kinase phosphorylates more than one response regulator). Although cross talk and cross-regulation have been reported for E. coli and Pseudomonas spp., they have not been studied in GAS (10, 30, 38, 42).

    The transcriptome data permit us to identify common themes regarding genes influenced by each of the four TCSs studied, thereby providing potential insight about functions of each TCS. The M5005_Spy_0680/0681 TCS apparently affects nucleotide and complex carbohydrate metabolism, especially during transition from logarithmic to stationary phase. Although speculative, it is possible that the observed down-regulation of the ccpA homolog in late exponential phase participates. The CcpA (carbon catabolite repression) protein is a major regulator of sugar transport and metabolism in gram-positive bacteria (43). Finally, we note that Virtaneva et al. (39) recently reported a significant correlation between expression of these genes and clinical pharyngitis in experimentally infected monkeys. This observation supports the idea that the M5005_Spy_0680/0681 TCS plays an important role in host-pathogen interaction.

    No distinct global theme was revealed by analysis of the transcripts influenced by the M5005_Spy_0784/0785 TCS. However, one apparent transcriptional unit may be directly regulated by this TCS. Four genes (M5005_Spy_0780-0783) encoding a putative mannose/fructose-PTS system were down-regulated over 200-fold in the late logarithmic and stationary growth phases in the mutant strain. These genes are located directly upstream of the two genes encoding this TCS (Fig. 1B). This fact, taken together with the extremely high level of transcript down-regulation observed in the mutant strain, suggests that it is a direct activator of this operon.

    Two observations suggest that the M5005_Spy_0830/0831 TCS is involved in malate metabolism. First, two genes (M5005_Spy_0832-0833) located directly downstream of this TCS were highly down-regulated in the mutant strain (see Table S1 in the supplemental material). These genes encode homologs of a malate-sodium symporter (M5005_Spy_0832) and an NAD-dependent malic enzyme (M5005_Spy_0833). Second, consistent with these findings, a distant Bacillus subtilis homolog of this TCS is involved in malate transport, inducible by external malate (37).

    The most prominent finding related to the M5005_Spy_1280/1281 TCS was the very large number of altered gene transcripts observed in the stationary phase. Only 16 transcripts were altered in the early logarithmic phase, whereas 473 were changed in the stationary phase, suggesting that this TCS functions as a major "switch" to change cell physiology late in growth. We find it remarkable that such a massive transcriptome change did not produce a significant difference in mouse virulence, an observation stressing the complexity of GAS pathogenesis.

    Most expression microarray studies have concentrated on analysis of single TCS mutants at one growth point rather than multiple strains and different growth phases (8, 22, 25, 29, 41). Our work adds to a growing list of reports describing expression array analysis of multiple TCSs in bacteria (14, 26, 27). One theme emerging from these investigations is that great variance exists in the number of genes directly or indirectly regulated by each TCS. For example, microarray analysis of mutations in all 32 E. coli TCSs identified three classes of transcriptional profiles based on the number of affected genes (minimal, moderate, or dramatic changes) (26). Our results mirror these findings. We found that three of the four GAS TCS mutants had moderate changes in transcriptome, whereas the transcript profile of M5005_Spy_1281 was dramatically changed. Given the very large number of transcripts altered in the M5005_Spy_1281 mutant strain, it is possible that it is a central actor in interfacing with multiple downstream regulators. Thus, inactivation of M5005_Spy_1281 would result in significant dysregulation of a large number of downstream genes. Consistent with this, expression of a very large number of transcription regulator genes was altered in the mutant strain (see Table S1 in the supplemental material).

    ACKNOWLEDGMENTS

    We thank S. F. Porcella, K. Virtaneva, and M. R. Graham for help with data analysis; R. Ireland and R. Larson for technical assistance; and R. A. Lempicki, C. Smith, and J. Yang for GeneChip hybridization. Partek, Inc., assisted with experimental design, and A. Whitney and F. DeLeo assisted with deposition of data in the GEO omnibus database.

    Supplemental material for this article may be found at http://iai.asm.org/.

    REFERENCES

    1. Affymetrix. 2005. GeneChip expression analysis technical manual. Affymetrix, Santa Clara, Calif. [Online.]

    2. Banks, D. J., S. F. Porcella, K. D. Barbian, S. B. Beres, L. E. Philips, J. M. Voyich, F. R. DeLeo, J. M. Martin, G. A. Somerville, and J. M. Musser. 2004. Progress toward characterization of the group A Streptococcus metagenome: complete genome sequence of a macrolide-resistant serotype M6 strain. J. Infect. Dis. 190:727-738.

    3. Beres, S. B., G. L. Sylva, K. D. Barbian, B. Lei, J. S. Hoff, N. D. Mammarella, M. Y. Liu, J. C. Smoot, S. F. Porcella, L. D. Parkins, D. S. Campbell, T. M. Smith, J. K. McCormick, D. Y. Leung, P. M. Schlievert, and J. M. Musser. 2002. Genome sequence of a serotype M3 strain of group A Streptococcus: phage-encoded toxins, the high-virulence phenotype, and clone emergence. Proc. Natl. Acad. Sci. USA 99:10078-10083.

    4. Edwards, J. A., G. W. Taylor, M. Ferguson, S. Murray, N. Rendell, A. Wrigley, Z. Bai, J. Boyle, J. S. Finney, A. Jones, H. H. Russell, C. Turner, J. Cohen, L. Faulkner, and S. Sriskandan. 2005. Specific C-terminal cleavage and inactivation of interleukin-8 by invasive disease isolates of Streptococcus pyogenes. J. Infect. Dis. 192:783-790.

    5. Federle, M. J., K. S. McIver, and J. R. Scott. 1999. A response regulator that represses transcription of several virulence operons in the group A streptococcus. J. Bacteriol. 181:3649-3657.

    6. Ferretti, J. J., W. M. McShan, D. Ajdic, D. J. Savic, G. Savic, K. Lyon, C. Primeaux, S. Sezate, A. N. Suvorov, S. Kenton, H. S. Lai, S. P. Lin, Y. Qian, H. G. Jia, F. Z. Najar, Q. Ren, H. Zhu, L. Song, J. White, X. Yuan, S. W. Clifton, B. A. Roe, and R. McLaughlin. 2001. Complete genome sequence of an M1 strain of Streptococcus pyogenes. Proc. Natl. Acad. Sci. USA 98:4658-4663.

    7. Graham, M. R., L. M. Smoot, C. A. Migliaccio, K. Virtaneva, D. E. Sturdevant, S. F. Porcella, M. J. Federle, G. J. Adams, J. R. Scott, and J. M. Musser. 2002. Virulence control in group A Streptococcus by a two-component gene regulatory system: global expression profiling and in vivo infection modeling. Proc. Natl. Acad. Sci. USA 99:13855-13860.

    8. Graham, M. R., K. Virtaneva, S. F. Porcella, W. T. Barry, B. B. Gowen, C. R. Johnson, F. A. Wright, and J. M. Musser. 2005. Group A Streptococcus transcriptome dynamics during growth in human blood reveals bacterial adaptive and survival strategies. Am. J. Pathol. 166:455-465.

    9. Green, M. N., S. Zhang, S. F. Porcella, M. J. Nagiec, K. D. Barbian, S. B. Beres, R. B. LeFebvre, and J. M. Musser. 2005. Genome sequence of a serotype M28 strain of group A streptococcus: potential new insights into puerperal sepsis and bacterial disease specificity. J. Infect. Dis. 192:760-770.

    10. Groisman, E. A. 2001. The pleiotropic two-component regulatory system PhoP-PhoQ. J. Bacteriol. 183:1835-1842.

    11. Gryllos, I., J. C. Levin, and M. R. Wessels. 2003. The CsrR/CsrS two-component system of group A Streptococcus responds to environmental Mg2+. Proc. Natl. Acad. Sci. USA 100:4227-4232.

    12. Hoe, N., K. Nakashima, D. Grigsby, X. Pan, S. J. Dou, S. Naidich, M. Garcia, E. Kahn, D. Bergmire-Sweat, and J. M. Musser. 1999. Rapid molecular genetic subtyping of serotype M1 group A Streptococcus strains. Emerg. Infect. Dis. 5:254-263.

    13. Hoe, N. P., K. Nakashima, S. Lukomski, D. Grigsby, M. Liu, P. Kordari, S. J. Dou, X. Pan, J. Vuopio-Varkila, S. Salmelinna, A. McGeer, D. E. Low, B. Schwartz, A. Schuchat, S. Naidich, L. D. De, Y. X. Fu, and J. M. Musser. 1999. Rapid selection of complement-inhibiting protein variants in group A Streptococcus epidemic waves. Nat. Med. 5:924-929.

    14. Kobayashi, K., M. Ogura, H. Yamaguchi, K. Yoshida, N. Ogasawara, T. Tanaka, and Y. Fujita. 2001. Comprehensive DNA microarray analysis of Bacillus subtilis two-component regulatory systems. J. Bacteriol. 183:7365-7370.

    15. Kreikemeyer, B., M. D. Boyle, B. A. Buttaro, M. Heinemann, and A. Podbielski. 2001. Group A streptococcal growth phase-associated virulence factor regulation by a novel operon (Fas) with homologies to two-component-type regulators requires a small RNA molecule. Mol. Microbiol. 39:392-406.

    16. Levin, J. C., and M. R. Wessels. 1998. Identification of csrR/csrS, a genetic locus that regulates hyaluronic acid capsule synthesis in group A Streptococcus. Mol. Microbiol. 30:209-219.

    17. Lipshutz, R. J., S. P. Fodor, T. R. Gingeras, and D. J. Lockhart. 1999. High density synthetic oligonucleotide arrays. Nat. Genet. 21:20-24.

    18. Lukomski, S., E. H. Burns, Jr., P. R. Wyde, A. Podbielski, J. Rurangirwa, D. K. Moore-Poveda, and J. M. Musser. 1998. Genetic inactivation of an extracellular cysteine protease (SpeB) expressed by Streptococcus pyogenes decreases resistance to phagocytosis and dissemination to organs. Infect. Immun. 66:771-776.

    19. Lukomski, S., N. P. Hoe, I. Abdi, J. Rurangirwa, P. Kordari, M. Liu, S. J. Dou, G. G. Adams, and J. M. Musser. 2000. Nonpolar inactivation of the hypervariable streptococcal inhibitor of complement gene (sic) in serotype M1 Streptococcus pyogenes significantly decreases mouse mucosal colonization. Infect. Immun. 68:535-542.

    20. Lukomski, S., C. A. Montgomery, J. Rurangirwa, R. S. Geske, J. P. Barrish, G. J. Adams, and J. M. Musser. 1999. Extracellular cysteine protease produced by Streptococcus pyogenes participates in the pathogenesis of invasive skin infection and dissemination in mice. Infect. Immun. 67:1779-1788.

    21. Lukomski, S., S. Sreevatsan, C. Amberg, W. Reichardt, M. Woischnik, A. Podbielski, and J. M. Musser. 1997. Inactivation of Streptococcus pyogenes extracellular cysteine protease significantly decreases mouse lethality of serotype M3 and M49 strains. J. Clin. Investig. 99:2574-2580.

    22. McCluskey, J., J. Hinds, S. Husain, A. Witney, and T. J. Mitchell. 2004. A two-component system that controls the expression of pneumococcal surface antigen A (PsaA) and regulates virulence and resistance to oxidative stress in Streptococcus pneumoniae. Mol. Microbiol. 51:1661-1675.

    23. Musser, J. M., A. R. Hauser, M. H. Kim, P. M. Schlievert, K. Nelson, and R. K. Selander. 1991. Streptococcus pyogenes causing toxic-shock-like syndrome and other invasive diseases: clonal diversity and pyrogenic exotoxin expression. Proc. Natl. Acad. Sci. USA 88:2668-2672.

    24. Nakagawa, I., K. Kurokawa, A. Yamashita, M. Nakata, Y. Tomiyasu, N. Okahashi, S. Kawabata, K. Yamazaki, T. Shiba, T. Yasunaga, H. Hayashi, M. Hattori, and S. Hamada. 2003. Genome sequence of an M3 strain of Streptococcus pyogenes reveals a large-scale genomic rearrangement in invasive strains and new insights into phage evolution. Genome Res. 13:1042-1055.

    25. Nishino, K., T. Honda, and A. Yamaguchi. 2005. Genome-wide analyses of Escherichia coli gene expression responsive to the BaeSR two-component regulatory system. J. Bacteriol. 187:1763-1772.

    26. Oshima, T., H. Aiba, Y. Masuda, S. Kanaya, M. Sugiura, B. L. Wanner, H. Mori, and T. Mizuno. 2002. Transcriptome analysis of all two-component regulatory system mutants of Escherichia coli K-12. Mol. Microbiol. 46:281-291.

    27. Paithoonrangsarid, K., M. A. Shoumskaya, Y. Kanesaki, S. Satoh, S. Tabata, D. A. Los, V. V. Zinchenko, H. Hayashi, M. Tanticharoen, I. Suzuki, and N. Murata. 2004. Five histidine kinases perceive osmotic stress and regulate distinct sets of genes in Synechocystis. J. Biol. Chem. 279:53078-53086.

    28. Parish, T., D. A. Smith, S. Kendall, N. Casali, G. J. Bancroft, and N. G. Stoker. 2003. Deletion of two-component regulatory systems increases the virulence of Mycobacterium tuberculosis. Infect. Immun. 71:1134-1140.

    29. Parish, T., D. A. Smith, G. Roberts, J. Betts, and N. G. Stoker. 2003. The senX3-regX3 two-component regulatory system of Mycobacterium tuberculosis is required for virulence. Microbiology 149:1423-1435.

    30. Ramos-Gonzalez, M. I., M. Olson, A. A. Gatenby, G. Mosqueda, M. Manzanera, M. J. Campos, S. Vichez, and J. L. Ramos. 2002. Cross-regulation between a novel two-component signal transduction system for catabolism of toluene in Pseudomonas mendocina and the TodST system from Pseudomonas putida. J. Bacteriol. 184:7062-7067.

    31. Ribardo, D. A., T. J. Lambert, and K. S. McIver. 2004. Role of Streptococcus pyogenes two-component response regulators in the temporal control of Mga and the Mga-regulated virulence gene emm. Infect. Immun. 72:3668-3673.

    32. Sanger Institute. 2005. Streptococcus pyogenes. Sanger Institute, Cambridge, United Kingdom. [Online.] http://www.sanger.ac.uk/Projects/S_pyogenes/.

    33. Smoot, J. C., K. D. Barbian, J. J. Van Gompel, L. M. Smoot, M. S. Chaussee, G. L. Sylva, D. E. Sturdevant, S. M. Ricklefs, S. F. Porcella, L. D. Parkins, S. B. Beres, D. S. Campbell, T. M. Smith, Q. Zhang, V. Kapur, J. A. Daly, L. G. Veasy, and J. M. Musser. 2002. Genome sequence and comparative microarray analysis of serotype M18 group A Streptococcus strains associated with acute rheumatic fever outbreaks. Proc. Natl. Acad. Sci. USA 99:4668-4673.

    34. Smoot, L. M., J. C. Smoot, M. R. Graham, G. A. Somerville, D. E. Sturdevant, C. A. Migliaccio, G. L. Sylva, and J. M. Musser. 2001. Global differential gene expression in response to growth temperature alteration in group A Streptococcus. Proc. Natl. Acad. Sci. USA 98:10416-10421.

    35. Sumby, P., K. D. Barbian, D. J. Gardner, A. R. Whitney, D. M. Welty, R. D. Long, J. R. Bailey, M. J. Parnell, N. P. Hoe, G. G. Adams, F. R. DeLeo, and J. M. Musser. 2005. Extracellular deoxyribonuclease made by group A Streptococcus assists pathogenesis by enhancing evasion of the innate immune response. Proc. Natl. Acad. Sci. USA 102:1679-1684.

    36. Sumby, P., S. F. Porcella, A. G. Madrigal, K. D. Barbian, K. Virtaneva, S. M. Ricklefs, D. E. Sturdevant, M. R. Graham, J. Vuopio-Varkila, N. P. Hoe, and J. M. Musser. 2005. Evolutionary origin and emergence of a highly successful clone of serotype M1 group A Streptococcus involved multiple horizontal gene transfer events. J. Infect. Dis. 192:771-782.

    37. Tanaka, K., K. Kobayashi, and N. Ogasawara. 2003. The Bacillus subtilis YufLM two-component system regulates the expression of the malate transporters MaeN (YufR) and YflS, and is essential for utilization of malate in minimal medium. Microbiology 149:2317-2329.

    38. Verhamme, D. T., J. C. Arents, P. W. Postma, W. Crielaard, and K. J. Hellingwerf. 2002. Investigation of in vivo cross-talk between key two-component systems of Escherichia coli. Microbiology 148:69-78.

    39. Virtaneva, K., S. F. Porcella, M. R. Graham, R. M. Ireland, C. A. Johnson, S. M. Ricklefs, I. Babar, L. D. Parkins, R. A. Romero, G. J. Corn, D. J. Gardner, J. R. Bailey, M. J. Parnell, and J. M. Musser. 2005. Longitudinal analysis of the group A Streptococcus transcriptome in experimental pharyngitis in cynomolgus macaques. Proc. Natl. Acad. Sci. USA 102:9014-9019.

    40. Voyich, J. M., K. R. Braughton, D. E. Sturdevant, C. Vuong, S. D. Kobayashi, S. F. Porcella, M. Otto, J. M. Musser, and F. R. DeLeo. 2004. Engagement of the pathogen survival response used by group A Streptococcus to avert destruction by innate host defense. J. Immunol. 173:1194-1201.

    41. Voyich, J. M., D. E. Sturdevant, K. R. Braughton, S. D. Kobayashi, B. Lei, K. Virtaneva, D. W. Dorward, J. M. Musser, and F. R. DeLeo. 2003. Genome-wide protective response used by group A Streptococcus to evade destruction by human polymorphonuclear leukocytes. Proc. Natl. Acad. Sci. USA 100:1996-2001.

    42. Wanner, B. L. 1992. Is cross regulation by phosphorylation of two-component response regulator proteins important in bacteria J. Bacteriol. 174:2053-2058.

    43. Warner, J. B., and J. S. Lolkema. 2003. CcpA-dependent carbon catabolite repression in bacteria. Microbiol. Mol. Biol. Rev. 67:475-490.(Izabela Sitkiewicz, and J)