CcpA-Mediated Repression of Streptolysin S Expression and Virulence in the Group A Streptococcus^▿,†

Bacterial strains and media.

The bacterial strains and plasmids used for this study are shown in Table 1. Streptococcus pyogenes MGAS5005 (covS) is a well-characterized invasive serotype M1 strain with an available genome sequence and is virulent in mice (51). GAS was cultured in Todd-Hewitt medium supplemented with 0.2% yeast extract (THY; Difco), and growth was assayed by measuring absorbance using a Klett-Summerson colorimeter. Chemically defined medium (CDM; 2×) was prepared according to the manufacturer's instructions (JRH Biosciences), followed by filter sterilization. Prior to use, freshly prepared sodium bicarbonate (44 mM) and l-cysteine (6.2 mM) were added, in addition to a sugar source at a final concentration of 0.25%. Escherichia coli strain DH5α (hsdR17 recA1 gyrA endA1 relA1) was used as the host strain for plasmid construction and was cultured in Luria-Bertani (LB) medium (EM Science). Antibiotics were used at the following concentrations: ampicillin at 100 μg/ml for E. coli, spectinomycin at 100 μg/ml for both E. coli and GAS, kanamycin at 50 μg/ml for E. coli and 300 μg/ml for GAS, and erythromycin at 500 μg/ml for E. coli and 1 μg/ml for GAS.

DNA manipulations.

Plasmid DNA was isolated using either a Wizard miniprep (Promega) or Midi/Maxi prep (Qiagen) purification system. DNA fragments were isolated from agarose gels by using a QIAquick gel extraction kit (Qiagen). GAS chromosomal DNA was isolated using previously described methods (3, 4). PCR for cloning was performed using Phusion high-fidelity polymerase (New England Biolabs), and reaction products were purified using the QIAquick PCR purification system (Qiagen). PCR for diagnostic assays was performed using Taq DNA polymerase (New England Biolabs). DNA sequencing was performed by GeneWiz, Inc.

Construction of the ΔccpA mutant MGAS5005.718.

SOE PCR was used to delete the ccpA gene. Briefly, the primers ccpA-PCRS#1 and ccpA-PCRS#2 (Table 2) were used to amplify a 1,005-bp upstream region containing the first six nucleotides of ccpA, a BglII site, and a 9-bp overlap with the second fragment at the 3′ end. The primers ccpA-PCRS#3 and ccpA-PCRS#4 (Table 2) were used to amplify an 1,115-bp downstream region containing the last 100 nucleotides of ccpA, with a BglII site at the 5′ end. These fragments were then combined as template DNA with the ccpA-PCRS#1 and ccpA-PCRS#4 primers (Table 2) to generate the deletion. The resulting product was blunt end ligated into EcoRV-digested pBluescript IIKS (−) to yield pKSM716. The nonpolar spectinomycin resistance gene was amplified from pSL60-1 (25) by use of the primers aad9-L2-bglII and aad9-R2-bglII (Table 2), digested with BglII, and ligated into BglII-digested pKSM716 to create pKSM717 (Table 1). The BamHI/XhoI ΔccpA aad9 fragment from pKSM717 was ligated with BamHI/XhoI-digested pJRS233 to yield pKSM718 (Table 1).

A ΔccpA mutant was created using temperature-sensitive allelic exchange as previously described (35). Mutants were screened for sensitivity to erythromycin and verified by PCR using the primers ccpA-L5 and aad9-R1 (Table 2) and by Southern blotting (Fig. 1B).

Construction of the ccpA-complementing plasmid pKSM719.

ccpA with its native promoter was amplified from MGAS5005, using the PCR primers PccpA-L1 and ccpAR1 (Table 2), and blunt end ligated into EcoRV-digested pJRS525 to create pKSM715 (Table 1). To produce a kanamycin-resistant complementing plasmid, the aad9 spectinomycin resistance gene from pJRS525 was removed by digestion with AflIII, the ends of the fragment were filled in, and it was further digested with SwaI. The aphA3 kanamycin resistance gene from puc4Ωkm2 was digested with SmaI and blunt end ligated into pJRS525 to yield pKSM201. The ccpAp-ccpA fragment from pKSM715 was cloned into pKSM201 by using PvuII/NcoI to create pKSM719 (Table 1).

Southern blot analysis.

Chromosomal DNA (7.5 μg) was digested with HindIII, separated in a 0.7% agarose gel, and transferred downward to a positively charged nylon membrane overnight under alkaline conditions, followed by UV cross-linking. The probe was PCR amplified using the primers Spy0515-L and -R (Table 2) and then radiolabeled with [α-³²P]dATP. The blot was hybridized for 2 h with 5 × 10⁶ cpm of the radiolabeled probe at 42°C, followed by two washes in 2× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) with 0.1% sodium dodecyl sulfate for 20 min at 42°C and then two washes in 0.1× SSC with 0.1% sodium dodecyl sulfate for 20 min. Blots were exposed to a phosphorimager cassette and visualized using a Storm 860 phosphorimager (GE Healthcare).

Northern blot analysis.

Northern blots of total RNA were performed using a NorthernMax protocol (Ambion) as previously described (38). Briefly, 1 to 10 μg of total RNA was separated in a formaldehyde-agarose gel, transferred to a positively charged nylon membrane, and UV cross-linked (Stratagene). The probe was amplified by PCR using the primers ccpA-L2 and R2 (Table 2) and then radiolabeled as previously described. Hybridization and scanning of the Northern blots were done as detailed above for Southern blots.

Murine infection models.

An overnight culture (5 ml) was used to inoculate 75 ml of THY and incubated static with appropriate antibiotics at 37°C until late logarithmic phase. Approximately 2 × 10⁷ CFU/ml, as determined by microscope counts and verified by plating for viable colonies, was used to infect 6- to 7-week-old female CD-1 mice (Charles River Laboratories). Mice were injected with 100 μl (2 × 10⁸ CFU) of the cell suspension by the i.p. route and were monitored as necessary for 72 h postinfection. Mice were euthanized by CO₂ asphyxiation upon signs of systemic morbidity (hunching, lethargy, and hind leg paralysis). Survival data were assessed by Kaplan-Meier survival analysis and tested for significance by the log rank test, using GraphPad Prism (GraphPad Software).

The invasive skin model of infection was performed as described previously (41). Six- to 7-week-old female CD-1 mice (Charles River Laboratories) were anesthetized and depilated over an ∼2-cm² area of the haunch with Nair (Carter Products, New York, NY), and 100 μl of a cell suspension (∼2 × 10⁸ CFU/mouse) was injected subcutaneously. Mice were monitored twice daily and were euthanized by CO₂ asphyxiation upon signs of morbidity. Lesion sizes (length by width) were measured at 72 h postinfection, with length determined at the longest point of the lesion. Lesion size data were analyzed using GraphPad Prism (GraphPad Software) and tested for significance using an unpaired two-tailed t test.

Microarray and real-time reverse transcription-PCR (RT-PCR) validation.

Microarray experiments were performed as previously described (39). Briefly, 10-μg samples of RNA from three biological replicates were isolated from MGAS5005 and the isogenic ΔccpA strain MGAS5005.718, using a Triton X-100 isolation protocol (48). RNAs were treated with DNase I and analyzed for quality on a formaldehyde-agarose gel. RNA samples were converted to cDNA with an amino-allyl UTP and were labeled with both Cy5 and Cy3 dyes, using a Cyscribe postlabeling kit (GE Healthcare), to allow for dye-swap experiments. Yields and incorporation of dye were determined using a Nanodrop ND-1000 instrument (Nanodrop Technologies). Equal volumes (25 μl) of labeled Cy5 cDNA and Cy3 cDNA were dried under vacuum, resuspended in 23.8 μl of distilled H₂O, and boiled for 5 min, followed by cooling on ice for 1 min. A 5× Hyb buffer (GE Healthcare) (17 μl) and formamide (27.2 μl) were added to the cDNA and applied to array slides under raised coverslips (Lifterslip, Inc). Microarray slides were hybridized at 50°C overnight in slide chambers (Array It). Slides were washed twice for 10 min each under the following buffer concentrations and temperatures: 6× SSPE-0.01% Tween 20 (1× SSPE is 0.18 M NaCl, 10 mM NaH₂PO₄, and 1 mM EDTA [pH 7.7]) at 50°C, 0.8× SSPE-0.001% Tween 20 at 50°C, and 0.8× SSPE at room temperature. Slides were scanned using a Genepix 4100A personal array scanner and GenePixPro 6.0 software (Axon Instruments).

Data from the output GenePix results file (gpr file) were analyzed using Acuity 4.0 software (Axon Instruments). Microarray data were normalized using the ratio of the means. Data sets were then generated by analyzing data points whose mean of the ratio (635/532) was ≥2.0 or ≤0.50, followed by removal of samples for which four of the six microarray hybridization experiments (67%) did not show significance. Array validation was carried out on 12 differentially regulated genes with real-time RT-PCR (see below) using real-time primer pairs (Table 2). Correlation coefficients for the arrays were determined by plotting the log value from the array (x) against the log value from real-time RT-PCR (y). An equation describing the line of best fit was determined, with the resulting R² value representing the fitness of the data, with higher correlations approaching an R² value of 1.

Real-time RT-PCR.

Briefly, 25 ng of DNase I-treated total RNA was isolated from each strain and added to SYBR green master mix (Applied Biosystems) containing 5 μg of each specific real-time primer for the one-step protocol (Table 2). The real-time RT-PCR experiments were completed using a Lightcycler 480 instrument (Roche), and levels presented represent ratios of wild-type (WT) to experimental values relative to the level of gyrA transcript.

SLS hemolysis assay.

Hemolysis assays were performed as previously described (37). Briefly, the WT, the ΔccpA mutant, and the complemented strain containing pKSM719 were grown in THY broth supplemented with 10% heat-inactivated horse serum. Samples were taken every hour for a total of 8 hours and immediately frozen at −80°C. Bacterial cells were pelleted, and a 1:10 dilution was made of the supernatant. Five hundred microliters of this dilution was added to an equal volume of 2.5% (vol/vol) difibrinated sheep red blood cells (RBC), which were washed three times with sterile phosphate-buffered saline, pH 7.4. This mixture was incubated at 37°C for 1 h and cleared by centrifugation at 3,000 × g. Supernatants were measured at 541 nm by use of a spectrophotometer (Molecular Dynamics) to determine release of hemoglobin by lysed RBC. Percent hemolysis was defined as follows: [(sample A − blank A)/(100% lysis A)] × 100. To assay for streptolysin O-mediated hemolytic activity, the SLS inhibitor trypan blue (13 μg/μl) was added to samples prior to incubation.

sagAp-luc construction (pKSM727).

The firefly luciferase gene (luc) was amplified from pLuc-MCS (Stratagene) by using the primers Luc-L and Luc-R (Table 2). The resulting fragment was blunt end ligated into EcoRV-digested pJRS525 to create pKSM720 (Table 1). Transformants were screened for orientation using the primers 1211 and Luc-R1 (Table 2). The sagAp promoter fragment was amplified using the primers PsagA-L1-B and PsagA-R1-X (Table 2), digested with BamHI and XhoI, and ligated into BglII/XhoI-digested pKSM720 to create pKSM727 (sagAp-luc) (Table 1).

Luciferase assay.

Luciferase assays were performed as follows. MGAS5005 transformed with pKSM727 was grown statically in 13 ml CDM supplemented with 0.25% of either glucose, sucrose, or a mixture of glucose and fructose and with the appropriate drug at 37°C. Upon reaching 30 Klett units, 500-μl samples were taken every 15 Klett units. Samples were pelleted, supernatant was discarded, and samples were placed at −20°C overnight. The luciferase assay was performed using a luciferase assay system (Promega). Pellets were resuspended in various amounts of 1× lysis buffer to normalize them to cell units according to the equation 4.5 = (x ml)(65 Klett units/2). The luciferase assay was read using a Centro XS³ LB 960 luminometer (Berthold Technologies), into which 50 μl of Luciferin-D reagent was directly injected.

Expression and purification of GAS HPr and HPr kinase.

Amino-terminal fusions of a six-His tag to both GAS HPr and HPr kinase were constructed as follows. A 261-bp region containing ptsH (HPr) and a 930-bp region containing ptsK (HPr kinase) were amplified from serotype M1 SF370 (Table 1) genomic DNA by using the primer pairs ptsHL2/ptsHR2 and HprK-NcoI L/HprK-XhoI R, respectively (Table 2). The resulting products were digested with NcoI and XhoI and ligated into NcoI/XhoI-digested pProEX-HTb to produce pKSM712 and pKSM713, respectively (Table 1). Following verification by PCR and DNA sequence analysis, pKSM712 and pKSM713 were transformed into E. coli BL21(DE3) Gold (Stratagene) for protein expression.

GAS His-HPr and His-HPr kinase were purified via Ni-nitrilotriacetic acid resin (Qiagen) under native conditions based on the manufacturer's protocol. Briefly, expression of protein was induced at an optical density at 600 nm of 0.6 nm for 4 h with 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside), and cell pellets were lysed in the presence of 1 mg/ml lysozyme and 1× Complete protease inhibitors (Roche), using a Branson sonicator (5 cycles of 30-s pulses at 50% duty cycle, with an output of 7.5). The protein concentration was determined for each fraction by use of protein assay reagent (Bio-Rad), using an Ultrospec 2100 spectrophotometer (GE Healthcare). Chosen fractions were dialyzed with two buffer changes in 4 liters of TKED buffer (100 mM Tris-HCl, 150 mM KCl, 1 mM EDTA, 0.1 mM dithiothreitol), and glycerol was added to 10% prior to storage of protein aliquots at −20°C.

To produce phosphorylated HPr (HPr-P), 20-μg aliquots of GAS His-HPr kinase and His-HPr were incubated in a reaction mixture containing 10 mM ATP, 20 mM Tris, 7.5 mM fructose 1,6-bisphosphate, 5 mM MgCl₂, and 1 mM dithiothreitol at 37°C for 15 min.

EMSA.

Electrophoretic mobility shift assays (EMSAs) were performed as previously described (2), using both PCR fragments and double-stranded oligonucleotides (ds-Oligo). For the PCR probe, a sagAp middle fragment was amplified using the primers M1_sagA cre-L and M1_sagA cre-R (Table 2), end labeled with [γ-³²P]ATP by use of T4 polynucleotide kinase (New England Biolabs), and purified by crush-and-soak elution. Twenty micromolar HPr-P was added to a constant amount (10 to 25 ng) of end-labeled sagAp probe, and increasing amounts of GAS His-CcpA (1 to 4 μM) were added. GAS His-CcpA was purified as described previously (2). Competition assays were performed by adding 500 ng of unlabeled probe (ccpAp, sagAp left, sagAp middle, or sagAp right) to binding reaction mixtures. After incubation for 30 min at 37°C, reaction products were separated in a 5% polyacrylamide-10% (vol/vol) glycerol gel at room temperature. Gels were dried under vacuum at 80°C for 1 h, exposed overnight to a phosphorimaging screen, and scanned using a Storm860 phosphorimager (Amersham Biosciences). Resulting data were analyzed with ImageQuant analysis software (version 5.0).

The oligonucleotide-based EMSA was performed as described above, except that ds-Oligo probes were generated by annealing 30-bp sense and antisense oligonucleotide pairs representing PccpACRE, PsagACRE, and randomly rearranged PccpACRE (scrambled). Annealed oligonucleotide probes were end labeled with [γ-³²P]ATP and purified across a Sephadex G-25 column (Roche). A constant amount of labeled ds-Oligo probe (ca. 1 to 5 ng) and increasing amounts of GAS His-CcpA (5 to 12.5 μM) were used in each reaction mixture. Competition assays were performed by the addition of 700 ng unlabeled ds-Oligo probes to binding reaction mixtures.

Construction of sagB strains.

To create non-SLS-producing strains, a polar mutation was made in sagB, the second gene in the sag operon. Briefly, a 500-bp fragment internal to sagB was amplified using the primers SagB-L and SagB-R (Table 2) and blunt end ligated into EcoRV-digested pJRS233 (35) to yield pKSM732. The resulting plasmid was electroporated into both WT and ΔccpA mutant strains by using a temperature-sensitive inactivation strategy as described previously (38). Strains were verified by loss of hemolysis on blood agar plates and by PCR using the primers 1211 and SagB-L or SagB-R (Table 2).

Microarray data accession number.

Array data have been submitted to the NCBI GEO database and are accessible through series number GSE11328.

A ΔccpA mutant shows increased virulence in mice.

To assess the role of CcpA in GAS pathogenesis, a ΔccpA mutant was constructed in the mouse-virulent M1 strain MGAS5005. The organization of the ccpA genomic region is highly conserved in the GAS (Fig. 1A). Similar to the case with other lactic acid bacteria, ccpA is divergently transcribed from the upstream pepQ XAA-Pro dipeptidase gene, with the predicted ccpAp promoter and cre present within the intergenic region, as previously described (2). However, GAS specifically possess two genes directly downstream of ccpA, encoding a glycosyl transferase (5005_Spy0425) and a glucosyl transferase (5005_Spy0426), followed by a Rho-independent transcriptional terminator (Fig. 1A). Although ccpA does not appear to be essential in GAS (24), attempts to inactivate ccpA using polar insertional strategies were unsuccessful. RT-PCR analysis found that ccpA and Spy0425 are transcriptionally linked (data not shown), suggesting that ccpA may be in an operon with a gene important for growth.

Therefore, an in-frame deletion of ccpA containing a nonpolar aad9 spectinomycin resistance cassette (25) was introduced into the genome of MGAS5005, and the mutation was verified by Southern blotting (Fig. 1B). The resulting ΔccpA mutant produced a small-colony phenotype on THY agar plates, which has been observed for ccpA mutants in other bacteria (21, 52). Although the ΔccpA strain produced a longer lag phase during growth in rich liquid medium, the mutant had no noticeable growth defects upon entering logarithmic phase compared to the parent strain MGAS5005 (data not shown). A complementing plasmid, pKSM719, containing ccpA under the control of its native promoter (ccpAp), resulted in elevated expression of ccpA transcripts, as demonstrated by Northern blotting (Fig. 1C).

To determine the role of CcpA in virulence, the ΔccpA mutant was assayed in two mouse models of GAS infection. In the first model of systemic infection, CD-1 mice were infected i.p. with either the WT MGAS5005 strain containing empty vector, the ΔccpA mutant with empty vector, or the ΔccpA mutant strain complemented with ccpA in trans. Surprisingly, infection of mice with the ΔccpA mutant led to a more rapid death than did infection with parental MGAS5005, with 90% lethality by 17 h postinfection (Fig. 2A). In addition, more than one-half of these mice exhibited severe hemorrhaging from several body sites, such as the rectum and mouth (data not shown). In comparison, only 50% of the MGAS5005-infected mice were dead at the end of 72 h, which is statistically significant (P < 0.0001). Overexpression of ccpA in the ΔccpA mutant complemented the increased virulence to slightly less than the WT level, linking CcpA to the observed hypervirulence phenotype. In a subcutaneous model of GAS skin infection, the ΔccpA mutant showed a significantly increased lesion size compared to those seen with both WT MGAS5005 and the complemented mutant (Fig. 2B). Interestingly, increased lethality was not observed with the ΔccpA strain, suggesting no enhancement of dissemination from the skin leading to systemic disease. These data strongly suggest that CcpA acts to repress virulence in the GAS, leading to increased pathogenesis in mouse models of both systemic and localized infections.

Determining the CcpA regulon in GAS.

To identify the CcpA-regulated genes that might be responsible for the increased virulence observed in mice, a transcriptome analysis of the ΔccpA mutant was undertaken (see Materials and Methods). WT MGAS5005 was compared to the isogenic ΔccpA mutant grown in rich THY medium with glucose to mid-log phase, a point in growth at which CcpA-mediated regulation would be expected to be strongest. A decrease (CcpA activation) or increase (CcpA repression) in transcript level in the mutant of >2-fold over three biological replicates was considered significant. Under these conditions, CcpA was found to regulate about 6% of the M1 MGAS5005 genome (124 genes), with the vast majority of regulated genes (116 genes [90%]) showing repression by CcpA (see Table S1 in the supplemental material). The microarray analysis was validated by real-time RT-PCR on 15 differentially regulated genes (Table 3), resulting in a strong correlation, with an r² value of 0.925.

Similar to studies on the CcpA regulon in other gram-positive bacteria (32, 52), CcpA repressed multiple operons important for nonglucose sugar utilization, including those encoding mannose, cellobiose, mannose/fructose, and β-glucosidase PTSs, as well as a maltose ABC transporter (see Table S1 in the supplemental material). It should be noted that not all alternative sugar utilization operons were regulated, suggesting that other mechanisms of CCR may exist in GAS (e.g., LacD1). The arginine deiminase operon (arcABC) was strongly repressed by CcpA, supporting previous studies showing that expression of the arc operon is inhibited by glucose, induced in stationary phase, and likely under CCR (5). A putative cre was previously identified upstream of arcA in an in silico search of the M1 GAS genome by use of the B. subtilis consensus cre (2). In fact, 76 of the 124 genes regulated by CcpA in the microarray study (61%) contained a predicted cre identified in that search or were associated with a cre present in the beginning of an operon. Ten percent of the CcpA-regulated genes encode transcriptional regulators, including the TCS response regulator TCS-5R (5005_Spy0785), which has been shown to regulate the adjacent mannose/fructose PTS operon (43), and the currently uncharacterized TCS-10R (5005_Spy1305), implying further indirect regulation of those genes lacking an identifiable cre.

As predicted by the infection studies, CcpA also regulates several genes important for GAS pathogenesis. In particular, the most highly CcpA-repressed locus in the array study represented the entire SLS operon (sagA to sagI) (Table 3; see Table S1 in the supplemental material). In addition to SLS being a well-characterized cytolysin and virulence factor (9, 31), the sagA locus also contains the pel regulatory RNA, which has been shown to regulate the expression of other virulence genes in GAS (27). However, we did not observe effects on emm, sic, or speB expression that would be predicted by altering pel expression (27). CcpA was also able to repress expression of spd, encoding a secreted DNase that contributes to the escape of GAS from the innate immune response (45). Finally, CcpA activated expression of the virulence regulator RivR (Ralp-4), which is both CovR repressed and can influence the expression of the Mga virulence regulon (40). Thus, CcpA appears to influence both carbon utilization and virulence gene expression in GAS.

Expression of SLS (sagA/pel) is catabolite repressed by CcpA.

The strong repression of sagA/pel by CcpA observed in the transcriptome analysis would predict a concomitant increase of SLS activity in the ΔccpA mutant during exponential phase. To investigate this, SLS-specific hemolytic activity was assayed using 2.5% defibrinated sheep RBC incubated with culture supernatants taken at 1-h time points during growth from WT MGAS5005 containing an empty vector, the ΔccpA mutant containing an empty vector, and the complemented ΔccpA mutant. SLS hemolytic activity showed a dramatic increase early in logarithmic phase for the ΔccpA mutant and remained elevated well into stationary phase (Fig. 3). In comparison, both WT MGAS5005 and the complemented ΔccpA strain exhibited very little SLS hemolytic activity during logarithmic phase, followed by a rapid increase to maximum levels at the transition to stationary phase (Fig. 3). In experiments using the SLS inhibitor trypan blue, no RBC lysis was seen for any of the samples, indicating that all hemolytic activity was due to SLS, not streptolysin O (data not shown). These results correlate with the CcpA-mediated repression of the sag operon during exponential growth observed with the microarray study.

The CcpA-mediated repression of sagA/pel transcription and SLS activity strongly suggests that sagA/pel expression is under CCR and is regulated directly by CcpA. Firefly luciferase (luc) has been used successfully in GAS as a transcriptional reporter and provides the unique ability to monitor promoter activity at various points during growth (36). Thus, a sagAp-luc reporter plasmid was introduced into WT MGAS5005, and the resulting strain was grown in CDM supplemented with either 0.25% glucose, sucrose (nonrepressing), or a mixture of glucose and fructose (nonrepressing) to assay for CCR of sagA/pel expression. The sagAp-luc construct showed low levels of luciferase activity across logarithmic phase that increased at stationary phase when cells were grown in glucose, indicating that CCR was occurring (Fig. 4). When the same strain was grown in a non-CCR-inducing sugar, such as sucrose, transcriptional activation of sagAp-luc was elevated earlier in logarithmic phase than that with glucose and reached higher overall levels. Importantly, when the strain was grown in a complex sugar environment consisting of equal parts glucose and the non-CCR-inducing sugar fructose, expression from sagAp demonstrated CCR. Early in growth, sagAp-luc expression mirrored that with glucose alone; however, it rapidly transitioned to the higher expression level characteristic of a non-CCR-inducing sugar when the glucose became depleted and fructose was utilized. These results indicate that expression of sagA/pel and SLS is under CcpA-mediated CCR.

CcpA binds directly to a cre in sagAp.

During a previous in silico analysis, a putative cre was identified in the promoter region of sagA/pel (sagAp), overlapping the predicted −35 region (2). Given our results so far, this strongly suggests that CcpA directly represses sagA production through binding to the sagAp cre. To investigate this possibility, an EMSA was performed using a radiolabeled sagAp probe (middle) containing the putative cre (Fig. 5A). Mobility of the sagAp middle probe was slowed in the presence of increasing amounts of purified GAS His-CcpA (1 to 3 μM) with 20 μM HPr-P-Ser. However, His-CcpA binding was comparable in the absence of HPr-P-Ser, suggesting that it was not required under these in vitro conditions (data not shown). The observed binding could be competed upon addition of cold sagAp middle probe but not with a sagAp probe that does not contain the cre (right probe), indicating that CcpA binds specifically to the middle probe (Fig. 5A and B). Interestingly, an upstream probe (left probe) that also lacks the predicted cre appears to compete slightly, suggesting that CcpA may more weakly bind to other regions of sagAp.

To further demonstrate the specificity of CcpA for the predicted sagAp cre, a 30-mer ds-Oligo probe encompassing the established ccpAp cre was used as previously described (2). Radiolabeled ccpAp ds-Oligo cre showed decreased mobility with increasing amounts (5 to 12.5 μM) of His-CcpA (Fig. 5C, lanes 1 to 5). The addition of 20 μM Hpr-P-Ser did not appear to enhance binding to oligonucleotide probes and was not utilized further (2; data not shown). Binding was competed upon addition of cold ccpAp ds-Oligo cre but not a scrambled version of the same cre probe (Fig. 5C, lanes 6 and 7), showing the specificity of His-CcpA binding. Importantly, cold ds-Oligo probes for the sagAp cre as well as cre within mgap and rivR also competed for binding of His-CcpA to ccpAp cre, to various degrees (Fig. 5C, lanes 8 to 10). Thus, CcpA binds directly to the cre present in sagAp and provides a mechanism for the observed CcpA-mediated CCR of SLS expression.

Role of SLS in CcpA-mediated repression of virulence.

Mutation of most genes in the sag operon, including sagB, leads to loss of SLS activity in GAS (9). To determine if the increased expression of SLS in the ΔccpA mutant contributes to the increased virulence seen in mice, the sagB gene was inactivated in both WT MGAS5005 and the ΔccpA mutant. Since the mutation of sagB is downstream of sagA/pel, it would be expected to inhibit SLS production independent of the pel transcript. Both the sagB single mutant and the sagB ΔccpA double mutant showed a complete loss of hemolysis on sheep blood agar plates compared to WT MGAS5005 (Fig. 6A). In addition, culture supernatants from both the sagB and sagB ΔccpA mutants did not exhibit hemolytic activity in the hemolysis assay (data not shown).

To assess the role of SLS in vivo, female CD-1 mice were infected i.p. with either WT MGAS5005, the ΔccpA mutant, the sagB single mutant, or the ΔccpA sagB double mutant at an average dose of 2.73 × 10⁷ CFU (Fig. 6B). The sagB single mutant had a small effect on virulence by the i.p. route of infection compared to WT MGAS5005, but this was not statistically significant (Fig. 6B). Published studies using the same route of infection also found a similar result (13). In contrast, inactivation of sagB in the ΔccpA mutant not only altered its hypervirulence phenotype but also resulted in significant attenuation (P < 0.0001) compared to the sagB single mutant (Fig. 6B). In addition, the sagB ΔccpA double mutant also showed significant attenuation (P < 0.0254) compared to WT MGAS5005. Therefore, the CcpA-mediated repression of virulence observed following i.p. infection is dependent on SLS production. Furthermore, in the absence of SLS, the loss of CcpA actually leads to attenuation of systemic virulence in GAS.

Defining the CcpA regulon of GAS.

In Lactococcus lactis and Bacillus subtilis, CcpA is a global regulator of gene expression, primarily affecting operons required for uptake and utilization of nonglucose sugar sources. However, in the oral pathogen Streptococcus mutans, CcpA was also able to regulate key virulence genes (1). To assess the CcpA regulon in the pathogenic GAS, we used a transcriptome analysis comparing the WT MGAS5005 strain with an isogenic ΔccpA mutant during exponential growth in rich medium, where glucose is present and CcpA activity would be expected to be highest. Our study identified 124 regulated genes (6% of the GAS genome) which were either up- or down-regulated at least twofold, with an added cutoff that four of six biological replicates with dye swaps must also be significant (see Table S1 in the supplemental material). This is comparable to the numbers of regulated genes observed in L. lactis (118 in early log phase and 86 in mid-log phase) and S. mutans (170 in mid-log phase) (1, 52). However, given that 148 CcpA-regulated genes of L. lactis were also identified during the transition from exponential to stationary-phase growth, the inclusion of later time points in our analysis would likely reveal even more of the GAS CcpA regulon.

The array analysis found that the majority of the GAS CcpA regulon (90%) is repressed during exponential growth phase, emphasizing that the primary function of CcpA is to down-regulate gene expression under high-glucose conditions. As expected, over 60% of the genes regulated by CcpA are involved in metabolism and carbohydrate transport, which corresponds to the results of studies with other gram-positive bacteria (1, 52). As might be expected from these results, detectable growth phenotypes, such as small colony size on plates and an increased lag phase in liquid media, could be observed in the CcpA mutant, although the growth rate was not significantly altered compared to that of the WT. Thus, CcpA appears to regulate carbon uptake and metabolism in response to glucose in GAS, and the inability to control this process does appear to have effects on GAS structure and growth.

Importantly, CcpA also regulates established virulence genes in GAS (see Table S1 in the supplemental material), including the secreted DNase B gene (spd) and the entire sag operon necessary for SLS synthesis and secretion. Several studies have indicated that expression of sagA/pel is tightly growth phase dependent, exhibiting low expression until transition into stationary phase (15, 27). Combined with the in silico analysis showing a putative cre within the sagA/pel promoter (2), these data strongly suggested that the sag operon was under CCR. This was confirmed here based on our transcriptome results, the sagAp-luc luciferase studies, and the SLS activity assays (Fig. 3 and 4). Furthermore, EMSA demonstrated specific binding of His-CcpA to the sagAp cre, indicating direct repression by CcpA. sagAp is also under direct repression by the CovRS TCS both in vitro and in vivo (11, 15, 19, 47). In fact, the sagAp cre characterized here overlaps the −35 region and falls within the region protected by CovR in DNase I footprints (15). Whether these two systems interact to control sag/pel expression is not clear. Interestingly, the MGAS5005 parental strain used here is a covS mutant that exhibits increased sagA production compared to that of WT strains and correlates with invasive potential in mice (47). Since both regulatory systems strongly repress SLS production in GAS, this suggests that this activity is tightly controlled during infection and expressed only under specific conditions.

CcpA represses the expression of TCS5 (5005_Spy1305/1306) and TCS10 (5005_Spy0784/0785) as well as activates expression of the RofA-like protein RivR (Table 3; see Table S1 in the supplemental material). TCS5 has been shown to positively regulate an adjacent mannose/fructose PTS operon that also appears as a repressed operon in our CcpA array data (Table 3) (43). RivR has been associated with positively influencing expression of the Mga virulence regulon in GAS (40). Recent studies found that CcpA activates transcription of mga by binding to a cre upstream of mgap (2). Interestingly, we did not find mga to be regulated significantly by CcpA in the microarray analysis. Since we have predicted that CcpA initiates mga expression very early in growth, prior to autoregulation, we may not see any regulation at the time points used here. Alternatively, since mga can be influenced by the pel regulatory RNA encoded within the sagA transcript (27), this may overcome the expected loss of mga expression in a CcpA mutant.

Determining a consensus GAS cre.

Of the 124 CcpA-regulated genes identified in our transcriptome analysis, 76 total genes contained 31 unique cre predicted in our in silico analysis, either within their promoter regions, within the genes themselves, or associated with a 5′ gene in the operon (2). Thus, 61% of the CcpA-regulated genes showed the potential for direct CcpA regulation through a cre identified using the B. subtilis consensus with one mismatch. Our binding studies have now found four cre that are bound specifically by CcpA in GAS, located in ccpAp, mgap, sagAp, and rivR (Fig. 5) (2). Alignment of these sites along with the 28 other unique GAS cre associated with CcpA-regulated genes produces a GAS consensus cre that is more flexible at five positions (positions 6, 7, 8, 9, and 13) and more specific at position 12 than the B. subtilis consensus (Fig. 7). This suggests that our initial screen was too strict and possibly missed potential sites. A new in silico search using the GAS consensus (Fig. 7) identified 13 new putative cre and 24 of 31 previously identified cre showing regulation on the microarray, with the remaining 7 being found with one mismatch. In addition, 11 more genes are associated with the newly described cre sites. Thus, 93/124 CcpA-regulated genes (75%) found in the array were associated with a cre by use of the GAS consensus. Overall, it appears that a GAS cre exhibits more flexibility at several positions than a B. subtilis cre does. However, this consensus will benefit from the study of more CcpA-cre interactions in GAS.

CcpA represses virulence in mice.

In this study, deletion of ccpA in the M1 strain MGAS5005 led to a surprisingly dramatic increase in virulence following both systemic (i.p.) and localized (subcutaneous) infections of mice. This was most evident in the i.p. model of systemic infection, where the animals succumbed faster than those infected with the WT and many exhibited evidence of severe hemorrhaging from various body sites. The hypervirulence phenotype in the GAS ΔccpA mutant was complemented beyond WT levels with ccpA overexpressed in trans, suggesting that the release of CCR had a significant effect on virulence gene expression in vivo. These results contrast with published studies of the closely related organism S. pneumoniae, where ΔccpA mutants were attenuated for virulence using three different mouse models of infection, including systemic (i.p.) infection, pneumonia, and nasopharyngeal colonization (16, 21). The reason for the observed attenuation in the ΔccpA mutant was suggested to be either reduced expression of cell wall proteins vital for metabolism in vivo or regulation of polysaccharide synthesis. However, we did not see comparable changes in GAS for surface proteins or capsule at mid-logarithmic phase in our array analysis.

SLS is a CcpA-repressed virulence factor during GAS systemic infection.

Among the most highly CcpA-repressed genes in the array analysis were sagA/pel and the entire sag operon, leading to an increase in SLS hemolytic activity (Table 3; Fig. 3 and 4). This raised the question of whether the increase in SLS production might be responsible for the hypervirulence seen in the mouse models of infection. Previous studies have shown that SLS contributes to the severity of localized necrotic lesions in mice (9, 13), which would help to explain this phenotype observed in the subcutaneous infection model (Fig. 2B). However, SLS-deficient mutants in an M5 GAS background did not show a significant virulence defect compared to the WT following i.p. infection of mice (13). The same thing was observed here, where a nonhemolytic sagB mutant of MGAS5005 did not show a significant difference from the parental strain following i.p. infection (Fig. 6B). In contrast, inactivation of SLS production in the ΔccpA mutant background (sagB ΔccpA) showed a dramatic reduction in virulence (P < 0.0001), abrogating the hypervirulence and resulting in an overall attenuation compared to either MGAS5005 or the sagB single mutant (Fig. 6B). This provides genetic evidence that the hypervirulent phenotype seen with the ΔccpA mutant during systemic infection is attributable to the increased expression of SLS in this strain. Furthermore, in the absence of SLS production, a ΔccpA mutation leads to attenuation.

The high level of CcpA-mediated repression of sagA might suggest that SLS has a role in secondary metabolism in the human host, where the function of the cytolysin in vivo may be to help release nutrients into the immediate environment surrounding the organism. This damage may have a cost of activating the host immune response, stimulating neutrophil migration to the site of infection and eliciting an inflammatory response. This would necessitate strict regulation of SLS expression, either through CcpA-mediated CCR or via CovRS repression. Thus, SLS expression is normally repressed in vivo, potentially by the presence of high levels of glucose in the peritoneum. Additional experiments will be necessary to determine if this also occurs during soft tissue infection at the skin.

During the final preparation of the manuscript, a study from Shelburne et al. was published which also described the role of CcpA in the pathogenesis of the M1 GAS strain MGAS5005 (42). In fact, the parental MGAS5005 strain used in our work was obtained directly from the same group. Notably, they analyzed the transcriptome of their ΔccpA mutant and found that CcpA strongly repressed expression of the sag operon and SLS activity, in addition to other virulence and regulatory genes. In the majority of cases, the results described here closely mirror their data and help to strongly validate both studies, with one significant exception. Shelburne et al. found that their ΔccpA mutant was attenuated following i.p. infection in female CD-1 mice, whereas our studies showed an increase in virulence in the identical background. They did not investigate the mechanism of CcpA-mediated attenuation in their study. Although the difference could be attributed to a secondary mutation obtained during mutagenesis of ccpA for one of the groups, successful complementation by both groups would appear to rule this out. Another possible explanation could be subtle differences in the execution of the i.p. infection model. Since we observed attenuation in a ΔccpA mutant in the absence of SLS production, this provides a potential mechanism for the differences seen in vivo. Regardless, the two studies clearly demonstrate for the first time a direct link between carbon metabolism and virulence in GAS. In contrast, our work has revealed a significant role for CcpA-mediated repression of the cytolysin SLS in the severity of GAS systemic infection that was not previously appreciated.