The CcpA regulon of Streptococcus suis reveals novel insights into the regulation of the streptococcal central carbon metabolism by binding of CcpA to two distinct binding motifs

Authors


  • Parts of the microarray data used in this manuscript had been published previously in Microbiology (Willenborg et al., 2011, 157: 1823–1833).

Summary

Streptococcus suis (S. suis) is a neglected zoonotic streptococcus causing fatal diseases in humans and in pigs. The transcriptional regulator CcpA (catabolite control protein A) is involved in the metabolic adaptation to different carbohydrate sources and virulence of S. suis and other pathogenic streptococci. In this study, we determined the DNA binding characteristics of CcpA and identified the CcpA regulon during growth of S. suis. Electrophoretic mobility shift analyses showed promiscuous DNA binding of CcpA to cognate cre sites in vitro. In contrast, sequencing of immunoprecipitated chromatin revealed two specific consensus motifs, a pseudo-palindromic cre motif (WWGAAARCGYTTTCWW) and a novel cre2 motif (TTTTYHWDHHWWTTTY), within the regulatory elements of the genes directly controlled by CcpA. Via these elements CcpA regulates expression of genes involved in carbohydrate uptake and conversion, and in addition in important metabolic pathways of the central carbon metabolism, like glycolysis, mixed-acid fermentation, and the fragmentary TCA cycle. Furthermore, our analyses provide evidence that CcpA regulates the genes of the central carbon metabolism by binding either the pseudo-palindromic cre motif or the cre2 motif in a HPr(Ser)∼P independent conformation.

Introduction

The catabolite control protein A (CcpA) is the major transcriptional regulator of the metabolic adaptation of Gram-positive bacteria to different carbohydrate sources. In homofermentative streptococci sugar components are transported into the bacterial cell by the phosphoenolpyruvate-dependent phosphotransferase system (PTS) (Vadeboncoeur and Pelletier, 1997). The PTS consists of the general components enzyme I (EI) and a phosphocarrier protein (HPr) that mediate internalization of all PTS sugars using a number of substrate-specific enzyme II (EII) complexes (Postma et al., 1993). HPr also acts a sensor of the energy state of the cell. In a nutrient-rich environment HPr becomes phosphorylated at a serine residue at position 46 [HPr-(Ser)∼P] by the HPr kinase/phosphorylase (HPrK/P) which activity depends on the cytosolic level of fructose-1,6-bisphosphate (FBP). Phosphorylated HPr-(Ser)∼P is a known co-regulator of CcpA. In many carbon catabolite controlled genes the CcpA–HPr-(Ser)∼P complex binds to a specific nucleotide sequence within the promoter or operator region, the so-called catabolite response element (cre), to execute its regulatory function (Deutscher et al., 2006). By this CcpA contributes to carbon catabolite repression (CCR) or carbon catabolite activation (CCA) of genes. The core cre consensus sequence TGWAARCGYTWNCW (N is any base, W is A or T, R is A or G, Y is C or T) has been determined for Bacillus subtilis, the model organism in studies on CCR in Gram-positive bacteria (Stulke and Hillen, 2000). However, this cre consensus sequence differs among Gram-positive bacteria and among individual genes (Zomer et al., 2007). Furthermore, the functionality of the cre site depends on the nucleotide composition of the pseudo-palindrome, and its localization relative to the transcriptional start point leads to activation, repression or ‘roadblock’ of transcription (Fujita, 2009).

From many studies in different Gram-positive bacteria it seems that the principal role of CcpA is to sense via HPr the presence of preferentially metabolized sugars, like glucose, and to suppress alternative energy providing mechanisms as long as the respective carbohydrate catabolic pathway is activated. Thereby CcpA contributes to the cellular energy homeostasis (Titgemeyer and Hillen, 2002). However, many recent studies showed that CcpA is also implicated in the regulation of different virulence-associated factors of streptococci. Hence, respective ccpA knockouts had significant effects on the virulence of the ccpA-deficient strains in mice infection models (Giammarinaro and Paton, 2002; Iyer et al., 2005; Abranches et al., 2008; Shelburne et al., 2008; 2010). Furthermore, microarray experiments pointed on a role of CcpA under conditions where carbohydrate sources are exhausted (Lulko et al., 2007; Carvalho et al., 2011).

Within the group of pathogenic streptococci Streptococcus suis (S. suis) is of particular interest as it is a zoonotic agent and able to induce sepsis and meningitis in humans with close contact to pigs or pig products (Clifton-Hadley and Alexander, 1980; Arends and Zanen, 1988; Chanter et al., 1993). S. suis has been reported to be most frequent cause of adult bacterial meningitis in Vietnam (Mai et al., 2008). In pigs S. suis is a common colonizer of the upper respiratory tract, but preferentially in young pigs S. suis is able to cross the epithelial barrier causing pathologies such as meningitis, septicaemia, arthritis, endocarditis, and bronchopneumonia. Noteworthy, pathogenicity of S. suis seems to be strongly associated with the availability of major central metabolic determinants such as arginine and glucose (Gruening et al., 2006; Fulde et al., 2011; Willenborg et al., 2011).

We have recently shown that CcpA is involved in the regulation of important virulence-associated factors under glucose rich growth conditions indicating a central role of CcpA for the biological fitness of S. suis (Willenborg et al., 2011). In the present study we correlated cDNA microarray analyses with the sequence information obtained from immunoprecipitated CcpA bound chromatin to define the primary regulon of CcpA and to dissect its role in S. suis metabolism. We found that CcpA participates in the regulation important central metabolic pathways by binding to two different cre motifs.

Results

Characterization of CcpA regulated genes during growth of S. suis

In the present study we were interested in the role of CcpA during growth of S. suis. We compared the transcriptome of the S. suis wild-type strain 10 and strain 10ΔccpA grown in THB broth to early stationary (stat) growth phase to gain more insight into the role of CcpA for gene regulation in a growth phase where glucose is completely consumed. Overall, 389 genes were found to be differentially expressed in strain 10ΔccpA as compared to wild-type, representing 19.8% of the S. suis genome (Table S2). Higher expression of 243 genes and lower expression of 146 suggested a repressing and activating regulatory role for CcpA at early stat growth. The number was considerably higher than in early exponential (exp) growth phase in which glucose is still present in the medium. Here, 259 genes were differentially expressed in the S. suis ccpA mutant strain, with 141 higher and 118 lower expressed genes (Table S3), as previously described (Willenborg et al., 2011). Noteworthy, in contrast to the exp growth phase gene expression levels of more than 10-fold higher were never observed in the array experiments with mRNA from stat growth phase (Table S4).

Among the regulated genes, expression of 86 was also affected at exp growth (Fig. 1A). COG clustering indicated that 36.1% of these genes were assigned to carbohydrate transport and metabolism. In addition, genes associated with transcriptional regulation (5.8%), amino acid and nucleotide metabolism (6.9%), chaperone functions (4.6%), and genes with yet unknown functions (25.6%) were affected in the mutant strain at both growth phases (Fig. 1B). Overall, 303 genes were not affected during exp growth which reveals that a subset of loci is influenced by CcpA in the early stat growth phase only. The majority of these genes encode for ribosomal proteins (13.5%) or could be assigned to carbohydrate (9.9%), amino acid (7.9%), or nucleotide (9.9%) metabolism respectively (Fig. 1B).

Figure 1.

Influence of CcpA knockout on global gene expression during growth of S. suis.

A. Venn diagram illustration of the number of significant differentially expressed genes during early exponential and early stationary growth of S. suis strain 10ΔccpA. Microarrays were validated by real-time qRT-PCR and subsequent correlation analysis of stationary grown wild-type strain 10 and strain 10ΔccpA (Fig. S1).

B. Summary of significantly differential expressed genes during exp and stat growth of S. suis strain 10ΔccpA and classification of clusters of orthologous groups (COG). C, energy production and conversion; D, cell cycle control, cell division; E, amino acid transport and metabolism; F, nucleotide transport and metabolism; G, carbohydrate transport and metabolism; H, coenzyme transport and metabolism; I, lipid transport and metabolism; J, translation; K, transcription; L, replication, recombination and repair; M, cell wall/membrane biogenesis; O, post-translational modification, protein turnover, chaperones; P, inorganic ion transport and metabolism; R, general function prediction only; S, function unknown; T, signal transduction mechanisms; V, defence mechanisms; [−], no prediction.

C. Heatmap depicting the influence of the CcpA knockout on the expression level of genes comprising the central carbon metabolism routes, like glycolysis, Leloir pathway, pentose phosphate pathway (PPP), tagatose metabolism and pyruvate metabolism. Mean fold changes (log2) of mRNA levels between wild-type and ccpA mutant are depicted as shades from red (lower mRNA level in strain 10ΔccpA) and blue (higher mRNA level). Genes are expressed as locus tag numbers or respective gene abbreviations.

When focusing on genes that are presumably subjected to CcpA mediated repression or activation in both growth phases, some of the CcpA repressed genes seemed to follow the classical CCR as they were either relieved from repression or, at least, their fold change gene expression increased during stat growth in the CcpA mutant strain. Most of these genes were assigned to carbohydrate metabolism, transporters or amino acid transport (Fig. 1B). Some prominent examples were genes of the arginine deiminase system (arcABC, SSU0580–583), the glycogen synthase cluster (glgCAB, SSU0870–0874), carbohydrate conversions (pgmA, SSU1230, SSU1169), or PTS for carbohydrate uptake of glucose/mannose (manLMN, SSU1583–1585), glycosaminoglycans (SSU1055–1057), beta-glucosides (SSU1855–1858, SSU1309) and trehalose (SSU0217). Notably, among the 18 putative PTS in S. suis only 4 were differentially expressed, suggesting that other CCR mechanisms may exist. Furthermore, expression of a sigma modulation factor (SSU0395) was strongly increased in strain 10ΔccpA giving evidence for indirect transcriptional effects of CcpA. Due to their significant lower expression levels in strain 10ΔccpA, potential targets of CcpA mediated CCA at early exp growth phase were the capsule synthesis cluster (cps2ABCEDFGHIJK, SSU0515–0526), a putative transcriptional antiterminator (SSU1310, licT), the lactate dehydrogenase (ldh, SSU0927), or virulence-associated factors like the opacity factor (ofs, SSU1474) and the surface antigen one (sao, SSU1201). The lower expression levels of mRNA such as those encoding for enolase (eno, SSU1320) and phosphoglycerate mutase (gpmA, SSU1451) suggested also a positive regulatory activity of CcpA in the stat growth phase. These data indicate that CcpA, in addition to its role as a carbon catabolite-dependent repressor, might directly interact as a positive regulator with promoters of genes responsible for a variety of cellular functions.

Next we were interested in the role of CcpA in the principal metabolic pathways of S. suis. According to the annotated genome S. suis possesses central constituents of the carbon metabolism such as glycolysis, pentose phosphate pathway (PPP), and pyruvate metabolism, but it lacks a complete tricarboxylic acid cycle. A heatmap of differential gene expression in strain 10ΔccpA assigning to glycolysis, PPP, Leloir pathway, tagatose and pyruvate metabolism is shown in Fig. 1C. The differential gene expression pattern indicates a major importance of CcpA for regulating glycolytic and glycogenic processes of the central metabolism. During exp growth, ccpA deficiency resulted in a higher expression of pgi (SSU 1836), encoding for the glucose-6-phosphate isomerase, the first enzyme of glycolysis. Lower expression of GpmA (SSU1451) and Eno (SSU1320) mRNA was detected at stat phase of bacterial growth indicating the influence of CcpA on the carbon flux through glycolysis. A contribution of CcpA in the regulation of glycolysis and Leloir pathway is indicated by the higher expression of the phosphoglucomutase (pgmA), galK, and galT at early exp growth phase. GalK and GalT convert galactose to Glc-1-P that in turn can be converted by PgmA or alternatively shuttled into glycogen synthesis. The genes encoding the enzymes for the glycogen synthesis (glgCAB) showed a 30-fold change of expression and belonged to the most pronounced regulated genes in strain 10ΔccpA indicating a major importance of CcpA for regulating glycogenic processes. Deletion of ccpA had no significant effect on any gene assigned to the PPP. On the other hand, mRNA expression levels of enzymes of pyruvate metabolism were different. Thus, during exp growth a lower expression of the lactate dehydrogenase (ldh) mRNA and higher mRNA expression of the pyruvate dehydrogenase complex (pdhABC), the pyruvate formate-lyase activating enzyme (pflC), and alcohol dehydrogenase (adhEP) were detected in strain 10ΔccpA. However, with the transition to the early stat growth phase this CcpA-dependent phenotype was abrogated. In summary, the expression profile of strain 10ΔccpA compared to its parental strain suggests a major role of CcpA in fine-tuning of the metabolic flow in glycolysis, different routes of galactose metabolism (Leloir, tagatose pathway) and mixed acid fermentation during growth of S. suis.

Characterization of CcpA binding to cre consensus sequences in vitro

The above data showed that CcpA is involved in the regulation of a defined subset of genes during early exp and early stat growth phase and indicated a contribution of CcpA to CCR and CCA in S. suis at both growth phases. Furthermore the regulatory activity of CcpA was more pronounced in early exp growth phase, as revealed by higher fold changes of expression of several genes between wild-type and ccpA mutant, and the high number of genes which expression was affected at early stat growth phase. These striking results prompted us to analyse the molecular properties of CcpA in S. suis more in detail.

In most Gram-positive bacteria phosphorylation of the cofactor HPr at the Ser46 residue [HPr(Ser)∼P] is required for its interaction with CcpA and enables and/or enhances DNA-binding activity of CcpA. Therefore, we first determined the levels of HPr∼P at the early exp and early stat growth phase by non-denaturing PAGE and immunoblotting with an α-HPr antiserum (Fig. 2A). In S. suis and other streptococci non-phosphorylated HPr is occurring in two isoforms due to N-terminal modification by a methionine aminopeptidase (Robitaille et al., 1991; Dubreuil et al., 1996). As shown in Fig. 2A, after native PAGE of protein extracts from the wild-type strain HPr was detected in four major protein bands at the early exp growth phase and two major protein bands at the early stat growth phase. This was also seen in protein extracts of the CcpA mutant strain. Pre-treatment of the extracts with λ-phosphatase, which removes phosphate groups from serine, threonine, and tyrosine residues, led to qualitative changes in the HPr protein banding pattern at the early exp growth phase. Hence, the protein bands of faster migrating HPr∼P forms were less intensively labelled. Heat treatment of the protein extracts before native PAGE for identifying thermo-stable phosphorylation revealed that almost half of the phosphorylated HPr proteins were HPr(Ser)∼P. Overall, these results indicate that during the early exp growth phase the HPr protein signal in S. suis comprises almost identical amounts of non-phosphorylated HPr, heat labile HPr(His)∼P, and CcpA interacting HPr(Ser)∼P respectively.

Figure 2.

CcpA interacts with HPr in vivo.

A. Analysis of the phosphorylation state of HPr by native PAGE. The wild-type (WT) and ccpA-deficient strain (ΔccpA) were harvested at early exp or early stat growth and lysed as described in Experimental procedures. Before native electrophoresis lysates were heated at 70°C for 10 min or pre-treated with λ-phosphatase overnight. Top: Immunoblots developed with an α-HPr antiserum. Bottom: Immunoblots of the same samples after denaturing SDS-PAGE.

B. Co-immunoprecipitation assay of HPr. Immunoblot analysis using an α-HPr antiserum of native (−) and cross-linked (+) cell lysates immunoprecipitated using an α-CcpA antiserum. The asterisk indicates the lane with the molecular weight marker and ΔC stands for lysate of the ccpA-deficient strain. Bottom panel shows a Coomassie blue stain of the input lysates as loading control.

Next, we investigated whether HPr interacts with CcpA in vivo. For this we performed co-immunoprecipitation experiments with cell lysates of S. suis wild-type strain 10 and, as a negative control, strain 10ΔccpA. Bacteria were grown to early exp or early stat growth phase and were subsequently fixed with EGS and formaldehyde or left untreated. CcpA complexes in the cell lysates were immunoprecipitated with an α-CcpA antiserum and protein-A/G agarose. Immune complexes were separated by SDS-PAGE and analysed for the presence of HPr after immunoblotting with an α-HPr antiserum. As shown in Fig. 2B, co-immunoprecipitation of HPr together with CcpA was detectable in cross-linked wild-type lysates. Markedly higher amounts of CcpA-bound HPr were detected in cell lysates of exp grown S. suis compared to stat grown S. suis.

The data shown in Fig. 2 suggested that more CcpA–HPr(Ser)∼P complexes are present in early exp growth phase, which corresponds to the data from other bacteria and indicates a contribution of CcpA to CCR and CCA in the early exp growth phase of S. suis. This was further emphasized by a pronounced influence of CcpA on gene expression observed in our microarray experiments. These data revealed the operons of the glycogen synthesis cluster (glgCAB) and arginine deiminase system (arcABC) to be regulated by CcpA mediated CCR, as they were markedly upregulated in strain 10ΔccpA during exp growth (Fig. 1C). To further elucidate the molecular characteristics of S. suis CcpA, we screened promoter/operator regions of these operons for putative CcpA binding sites (cre sites) according to the B. subtilis cre consensus sequence with the Virtual Footprint software (Munch et al., 2005). By this we were able to identify a putative cre site with highest prediction scores for the arcABC operon (Score: 10.56, position −130 relative to ATG) and for the glgCAB operon (Score: 11.08, position −71 relative to ATG) as depicted in Fig. 3A.

Figure 3.

CcpA binds to pseudo-palindromic cre sites in vitro.

A. Nucleotide sequences of oligonucleotides harbouring the pseudo-palindromic cre sites of the glgC promoter (glgC) and arcA promoter (arcA), respectively, used for EMSA. Mutations within point mutated oligonucleotides (glgCMUT; arcAMUT) are underlined.

B–E. EMSA analyses of wild-type (WT) or strain 10ΔccpA (ΔC) lysates (B–D), or recombinant CcpA (E). EMSA was carried out with 200 ng recombinant CcpA or 10 μg of bacterial cell lysate and 20 000 cpm of radiolabelled oligonucleotide probes. Competition experiments were performed with 200-fold molar excess of the corresponding un-labelled oligonucleotide (arcA, or glgC), the corresponding point-mutated (arcAMUT or glgCMUT), or scramble oligonucleotide. Immunoshift assays were performed with an α-CcpA antiserum, α-HPr antiserum or preimmune serum.

Following, double-stranded oligonucleotides with the identified pseudo-palindromic cre sites, glgC and arcA, as well as the respective oligonucleotides with point mutated cre sites (G9 to A and C14 to T) glgCMUT and arcAMUT (Fig. 3A) were analysed by electrophoretic gel shift assays (EMSA) for their possible interaction with CcpA. EMSA results shown in Fig. 3B and C clearly demonstrated binding activity of CcpA proteins to the radioactively labelled glgC and arcA oligonucleotides (lane 1), which could be competed with a 200-fold excess of the respective unlabelled oligonucleotides (lane 2), but not by the respective unlabelled mutated oligonucleotides (lane 3). Addition of an α-CcpA antiserum to the binding reaction before native gel electrophoresis prevented protein DNA complex formation revealing CcpA as major binding protein of both cre sites (lane 5). In contrast, addition of an α-HPr antiserum did not influence the CcpA–cre complex (lane 6). This was surprising for the cell extracts from early exp growth phase, but it might be a consequence of the low amounts of HPr bound CcpA molecules in the cell lysates as indicated in Fig. 2B. Nevertheless, the data demonstrate that under our conditions CcpA is actively binding in the absence of HPr. The latter is underlined by the binding activity similar to that in cell lysates from exp grown bacteria observed when the oligonucleotides were incubated with cell lysates from S. suis at early stat growth phase (lane 8). The EMSA in Fig. 3D demonstrates that the glgC and arcA oligonucleotides competed mutually. EMSA with recombinant CcpA and radioactively labelled arcA oligonucleotide confirmed that the binding capacity of S. suis CcpA to the cre site in vitro is independent of other cofactors (Fig. 3E).

Overall, the EMSA experiments demonstrate that in vitro CcpA of S. suis binds specifically to a pseudo-palindromic cre consensus sequence present within the arcABC and glgCAB promoter, irrespectively of the growth phase and of the interaction with HPr.

Dissection of CcpA binding in vivo

From the above data we assumed that CcpA binding to different cre sites in vitro was independent of the bacterial growth phase or other cofactors. In order to analyse the relevance of the pseudo-palindromic cre sites of the glgCAB- and arcABC-promoter–operator regions for CCR in vivo we cloned the non-coding promoter–operator regions of the glgCAB operon or arcABC operon with and without point mutated cre sites into the expression vector pGA14-gfp and transformed it into S. suis wild-type strain 10 and the CcpA mutant strain 10ΔccpA. To monitor CCR, bacteria were grown in TY medium containing glucose or galactose to exp growth phase, and then episomal reporter gene expression was determined by fluorescence measurement. As shown in Fig. 4A the arcA promoter was strongly subjected to glucose mediated CCR whereas galactose had no repressive effects on promoter activity. In contrast, the glgC promoter was subjected to both, glucose and galactose mediated CCR. Most interestingly, the point mutation of the cre site of the glgC promoter resulted in a significant (***P < 0.001) higher promoter activity compared to that of 10::glgC-gfp, whereas the point mutation of the predicted cre site of the arcA operator did not lead to a significant relief from CCR. Furthermore, the glgC, but not the arcA promoter activity was increased significantly (***P < 0.001) in strain 10ΔccpA::glgC-gfp indicating that CCR of the glgC gene but not of the arcA was directly attributed to CcpA.

Figure 4.

glgC but not arcA is directly regulated by CcpA in vivo.

A. GFP reporter assay. Reporter plasmids carrying the GFP under control of the arcA or the glgC promoter with (10::arcAMUT-gfp; 10::glgCMUT-gfp) or without point mutated cre sites (10::arcA-gfp; 10::glgC-gfp) were transformed in S. suis wild-type strain 10 and strain 10ΔccpA (10ΔccpA::arcA-gfp; 10ΔccpA::arcAMUT-gfp; 10ΔccpA::glgC-gfp; 10ΔccpA::glgCMUT-gfp). Bars represent the relative fluorescence units (RFU) after normalization to the values obtained for strain 10 carrying the promoterless gfp construct (10::gfp). Experiments were carried out in triplicate and repeated at least thrice.

B. ChIP assays were performed for wild-type strain 10 (WT) and strain 10ΔccpA (ΔccpA) grown to early exponential (exp) and early stationary phase (stat). ChIP lysate chromatin was precipitated using either α-CcpA or preimmune serum. ChIP-DNA was quantified along a serial dilution of defined amounts of Input-DNA by real-time qPCR using primer pairs for the respective promoter region of interest. The graphs represent means and standard deviations of immunoprecipitated DNA from three independent experiments. Significance was calculated by two-sample t-test.

In order to dissect CcpA binding to the promoter–operator regions of the glgCAB and arcABC operon on the chromatin level in vivo we performed chromatin immunoprecipitation followed by quantitative real-time PCR. For this, cross-linked bacterial chromatin from S. suis grown to the early exp and the early stat growth phase, and from strain 10ΔccpA as a negative control, was immunoprecipitated with an α-CcpA antiserum. The specificity of our ChIP assays in general and in particular for CcpA has been proven by the experiments shown in supplementary data (Figs S2 and S3A). As shown in Fig. 4B, after ChIP with the CcpA-specific antiserum we were able to precipitate and amplify DNA fragments of the glgC promoter region in the lysates from both growth phases. Although our cDNA microarray experiments suggested CcpA mediated CCR for glgC in the early exp growth phase and only moderate relevance for glgC expression in the early stat growth phase we observed higher binding activity for CcpA in the early stat growth phase. In agreement with our reporter studies, neither during early exp nor during early stat growth arcABC promoter DNA could be precipitated and amplified in significant amounts demonstrating that CcpA is not able to bind the arcABC promoter in vivo.

Taken together, these results suggest constitutive binding activity of CcpA to the glgC promoter region in vivo and that the regulatory capacity of CcpA most likely depends on cofactors such as HPr. Furthermore they indicate that in vivo CcpA binding to its cre motif depends on the context of the regulatory elements of the endogenous promoter.

Distribution of CcpA bound chromatin loci in vivo

In order to define the complete CcpA regulon in vivo, chromatin of early exp and early stat grown S. suis was immunoprecipitated with the α-CcpA antiserum and then analysed by deep sequencing (ChIP-Seq). ChIP-seq data were normalized, and significantly enriched regions, exemplarily illustrated by the peak resulting from the enriched glgC promoter DNA (Fig. S3B), were identified as described in Experimental procedures. Based on the peak detection algorithm, we identified 115 peaks corresponding to CcpA bound DNA loci (following designated as CBD) during exp and stat growth (Table 1). These peaks were significantly enriched when compared to a ChIP with the α-CcpA serum of 10ΔccpA chromatin or with a preimmune serum and wild-type chromatin (data not shown).

Table 1. CcpA bound DNA loci identified by ChIP-seq at the early exponential and early stationary growth phase of S. suis
Locus tagaGeneDescriptionCOGChIP-seq (exp)Fold change array (exp)bChIP-seq (stat)Fold change array (stat)bGenome positioncEnriched regiondcre motifeLocationfcre2 motifeLocation
  1. a. Genes are categorized based on COG categories. SSU numbers are the open reading frames based on the S. suis strain P 1/7 genome annotation (Holden et al., 2009). Where a number is underlined, the description is specific to the underlined SSU gene designation. Asterisks (*) indicate first gene of a regulated operon.
  2. b. Fold change gene expression is given for the first gene of a regulated gene/operon as positive (+) or negative (−) value representing either upregulation or downregulation in strain 10ΔccpA respectively. Significant genes are shown for the lowest false discovery rate of q = 0.5. NS: no significant difference between wild-type and mutant.
  3. c. Significant enriched peaks (including peak length) calculated by the CisGenome software were mapped to the genome position of S. suis strain P 1/7.
  4. d. Prom: significant peak enrichment in non-coding promoter–operator region; ORF: significant peak enrichment in protein encoding open reading frame.
  5. e. Listed motifs were detected after using the MEME computed cre or cre2 consensus motif and subsequent application using FIMO with a cut-off value of P < 10−5. NS: no significant motif found below the cut-off value.
  6. f. The position of the outermost end of a cre or cre2 site relative to the respective translation start point of the (first) gene. NS: no significant motif found below the cut-off value.
SSU0191pflBFormate acetyltransferaseCYes−2.41Yes−7.89191181 … 191554 (317)PromTTGAAAACACTTTCTA−85GTTTTTCATTTATTTTCAAGT−118
SSU0261adhEAldehyde-alcohol dehydrogenase 2CYes2.49YesNS266661 … 267111 (451)PromATGGAAGCGTTTTCTT−272NSNS
SSU0675*glpKGlycerol kinaseCYes2.11YesNS700602 … 700908 (307)PromATGGAAGCGTTTTCAA−46NSNS
SSU0927ldhl-lactate dehydrogenaseCYes−2.81Yes−3.63955907 … 956243 (337)PromATGTAAACATTTTCAA−114NSNS
SSU1042*acnAAconitate hydrataseCYesNSYesNS1063766 … 1064285 (520)ORFNSNSNSNS
SSU1637*pdhAPutative pyruvate dehydrogenase E1 componentCYes3.52YesNS1653279 … 1653627 (349)PromATTGAACCGTTTTCTT−82GGTTTTCACAAAAATTTCTGT−106
SSU0010 Septum formation initiator proteinDYesNSYesNS9778 … 10384 (607)ORFNSNSCTTTTTCTATCCTTTTTTCAC1
SSU0229/0230 ABC transporter ATP-binding proteinEYesNSYesNS235917 … 236466 (550)ORFNSNSCTTTTTCAAATCCTTTCCCTT181
SSU0537 TransferaseENo−3.1YesNS572472 … 572751 (280)ORFNSNSNSNS
SSU1194 Putative amino acid ABC transporterENoNSYes2.261216000 … 1216349 (350)ORFNSNSTTATTTCAAAGATTTTCCAAT48
SSU1364 Branched-chain amino acid ABC transporterEYes−3.16YesNS1390866 … 1391106 (241)PromNSNSNSNS
SSU1611 AspartokinaseEYesNSYesNS1617650 … 1618121 (472)PromTTGAAAACTTTTTCTA49GTTTTTCTGAAATTTTCATAC−110
SSU1741leuA2-Isopropylmalate synthaseENoNSYesNS1759549 … 1760008 (460)PromNSNSATTTTTCAGAAAATTTCCATT−287
SSU1682alsSAcetolactate synthase large subunitEHNoNSYes2.421698145 … 1698628 (484)PromNSNSAATTTTCAGAAAATTTCTTGA−61
SSU0315 L-asparaginaseEJYesNSYesNS335345 … 335534 (190)ORFNSNSTTTTTTCATAAATTTTCAGTT9
SSU0875 Extracellular amino acid-binding proteinETYesNSYes2.01902564 … 902909 (346)ORFNSNSCTTTTTCTTTAATTTTATCTA85
SSU0681 Putative oligopeptidaseFYes7.86Yes2.26708423 … 708792 (370)PromNSNSCTTTTTCATAAATTTTGCCTT−98
SSU0735*pyrRPyrR bifunctional regulatory proteinFNoNSYes3.29766085 … 766535 (451)PromNSNSTCTTTTCATAAAAATTCCAAT−380
SSU0778 Thymidylate synthaseFNoNSYesNS812238 … 812535 (298)ORFNSNSNSNS
SSU0789tdkThymidine kinaseFNoNSYesNS821883 … 822117 (235)ORFNSNSNSNS
SSU0935cddCytidine deaminaseFNo3.37YesNS962850 … 963249 (400)ORFNSNSNSNS
SSU0937deoAThymidine phosphorylaseFNo5.62YesNS963750 … 964249 (500)ORFNSNSTGTTTTCCGAAATATTTTCAT1178
SSU1355 Surface-anchored 5′-nucleotidaseFYesNSYes−2.081382163 … 1382532 (370)PromNSNSNSNS
SSU1758/1759 Adenylosuccinate synthaseFNoNSYesNS1779030 … 1779501 (472)PromNSNSGATTTTCAGTACATTTTTCAC−87
SSU1901 NucleotidaseFNoNSYesNS1933800 … 1933999 (200)PromNSNSNSNS
SSU0249glaGlycerol facilitator-aquaporinGYes−7.89No−2.39254440 … 255944(505)PromGAGTAAACGTTTTCTT−109NSNS
SSU0312fbaFructose-bisphosphate aldolaseGNoNSYes−2.39324797 … 325040 (244)PromNSNSNSNS
SSU0329galKGalactokinaseGYesNSNoNS347777 … 348113 (337)PromTTGTAATCGTTTTCTT−79NSNS
SSU0483tpiATriosephosphate isomeraseGYesNSYes−2.35516250 … 515449 (200)ORFNSNSNSNS
SSU0644pgmAPutative phosphomannomutaseGYes5.44YesNS667516 … 667813 (316)PromTATTAAACGCTTTCAT−41NSNS
SSU0870*glgCGlucose-1-phosphate adenylyltransferaseGYes29.46Yes3.7896032 … 896251 (220)PromAAGGAAACGCTTTCAT−71TATTTTCATTTTTCTTTTCAA−106
SSU0890 Aldose 1-epimeraseGYesNSYesNS924300 … 924449 (150)ORFNSNSNSNS
SSU0892*lacEPutative lactose-specific PTS. IIBC componentGNo2.4Yes3.44926592 … 926916 (325)ORFNSNSCTCTTTCAAGCTATTTTCAAC1293
SSU1169 Beta-fructofuranosidaseGNo4.9Yes2.631189753 … 1189927 (175)PromNSNSNSNS
SSU1230 N-acetylmannosamine-6-phosphate 2-epimeraseGYes3.12Yes3.891259437 … 1260310 (874)PromGAGATAGCGCTTACAT−44NSNS
SSU1265glgPPutative glycogen phosphorylaseGYes8.71Yes2.611296030 … 1296450 (421)PromTTGTAAACGTTTTCTT−71NSNS
SSU1320enoEnolaseGNo−2.81Yes−5.171352381 … 1352759 (379)PromNSNSTTTTTTCTTGAAAAATCCCTT−76
SSU1368dexBGlucan 1,6-alpha-glucosidaseGNoNSYes−2.351394750 … 1394999 (250)ORFNSNSTATTCTCAAGCATTTTTTCAA1630
SSU1368dexBGlucan 1,6-alpha-glucosidaseGYesNSYes−2.351392929 … 1393241 (313)PromNSNSNSNS
SSU1369gtfASucrose phosphorylaseGYesNSYesNS1395900 … 1396449 (550)PromTTGTAAGCCTTTTCTT−29NSNS
SSU1836pgiGlucose-6-phosphate isomeraseGYes2.09YesNS1861249 … 1861927 (679)PromNSNSTTTTTTCAAGAAATTTATCAA−49
SSU1849apuASurface-anchored amylopullulanaseGYesNSNoNS1880159 … 1880576 (418)PromAAGAAAACGTTTGCAA−104NSNS
SSU1915*malXPutative maltose/maltodextrin-binding proteinGYes5.49No−3.181947142 … 1947754 (613)PromATAAAAACGCTTGCAA−111NSNS
SSU1927 Beta-glucosidaseGYesNSNoNS1963491 … 1963815 (325)PromATTAAAACGCTTTCTT−70NSNS
SSU1929*bglAPutative beta-glucosidaseGYes−2.04NoNS1967196 … 1967757 (562)PromATGAAAACCCTTTCTT−96NSNS
SSU0282 Hypothetical proteinHYesNSYesNS295123 … 295357 (235)PromNSNSNSNS
SSU1830 5-Formyltetrahydrofolate cyclo-ligase familyHYesNSYes−2.061855933 … 1856416 (484)PromNSNSNSNS
SSU1954 CDP-diacylglycerol-3-P-phosphatidyltransferaseIYesNSYes2.471989013 … 1989526 (514)ORFNSNSNSNS
SSU0080rplN50S ribosomal protein L14JNoNSYes2.2377558 … 77792 (235)ORFNSNSNSNS
SSU0395yvyDSigma 54 modulation proteinJYes26.82YesNS419933 … 420314 (382)PromTTGCAACCGTTTTCTT−79NSNS
SSU0482tufAElongation factor Tu (EF-Tu)JYesNSYesNS513478 … 513892 (415)PromTTGCAAGCATTTTCTA−111NSNS
SSU1179 Peptide chain release factor 1JNoNSYesNS1198773 … 1199184 (412)ORFNSNSTTGTTTCATGACATTTTTGGC1176
SSU0226 Transcriptional regulatorKNoNSYesNS230350 … 230549 (200)ORFNSNSCATTTTCATAACCTTTCTGTA3
SSU0302 Transcription regulation proteinKYesNSNoNS552600 … 553149 (550)PromNSNSNSNS
SSU0355 GntR family regulatory proteinKYesNSNoNS380000 … 380449 (450)PromGCGTAACCGTTTTCTT−177GCTTTTCCGTAATATTTTAAA−120
SSU0368 Cold shock proteinKNo2.12YesNS392400 … 392849 (450)PromNSNSNSNS
SSU0613 Conserved hypothetical proteinKYes5YesNS638828 … 639341 (513)PromTCGAAAACGATTTCAA−61NSNS
SSU0908 Phage repressor-like proteinKNo−2.71YesNS937950 … 938099 (150)ORFNSNSGCTTTTCAATTTGTTTTCAGT891
SSU1172 Binding-protein-dependent transport systemKYesNSNoNS1193400 … 1193549 (150)PromCAGAAAACGGTTACAA−93NSNS
SSU1202ccpACatabolite control protein AKYesNSYes−2.641226201 … 1226603 (403)PromTTGAAACCACTTTCAA−91NSNS
SSU1826/1827 MarR family regulatory proteinKNoNSYesNS1852700 … 1852949 (250)PromNSNSTCTTTTCTGAAAATATTTTGT−102
SSU1850 LacI family regulatory proteinKYesNSNoNS1881465 … 1881663 (199)PromGTGCAAACGTTTTCTT−77NSNS
SSU0707addBATP-dependent exonuclease subunit BLYesNSYesNS738176 … 738617 (442)ORFNSNSNSNS
SSU1629ssbSingle-strand binding protein (SSB)LNoNSYes2.071642540 … 1642846 (307)ORFNSNSNSNS
SSU0515*cps2AIntegral membrane regulatory protein WzgMYes−2.95NoNS552536 … 553079 (544)PromAAGATAACGTTTTCCA−95NSNS
SSU0535neuBN-acetylneuraminic acid synthaseMNo−2.55YesNS570933 … 571311 (379)ORFNSNSNSNS
SSU1186penAPenicillin-binding protein 2bMYesNSNo2.671208700 … 1208949 (250)ORFNSNSNSNS
SSU1665 d-alanyl-d-alanine carboxypeptidaseMYesNSYesNS1678806 … 1679334 (529)PromNSNSTTTTTTCCAAATTTTTCTCAC−68
SSU1889 Accessory pilus subunit protein/pilusMYesNSYesNS1921613 … 1922063 (451)ORFNSNSTATTTTCAAGTTCATTTTCAT2630
SSU0159/0160 Glycoprotease family proteinOYesNSYesNS150750 … 151197 (448)PromNSNSNSNS
SSU0757 Cell envelope proteinaseONoNSYesNS789650 … 789799 (150)PromNSNSCTTTTTCATAACTTTTCCTCC−13
SSU0806 Plasmid replication proteinONoNSYesNS837102 … 837450 (348)ORFNSNSCATTTTCTTTCCAATTTTCAA607
SSU0806 Plasmid replication proteinONoNSYesNS837800 … 837998 (198)ORFNSNSCCTTTTCTGACATTTTTCTCT14
SSU1300 ProteaseONoNSYesNS1331344 … 1331713 (370)PromNSNSNSNS
SSU1480pflCPyruvate formate-lyase activating enzymeONo7.96YesNS1496952 … 1497384 (433)PromATAGAAGCGTTTTCAT−56NSNS
SSU1821radAPutative DNA repair proteinONoNSYes2.441846150 … 1846399 (250)ORFNSNSNSNS
SSU1968htrASerine proteaseONoNSYes−2.62005050 … 2005549 (500)PromNSNSACTTTTCTTTCAATTTTCCAA−217
SSU0953pstSPhosphate ABC transporterPNoNSYesNS978905 … 979283 (379)PromNSNSTTTTTTCATTTCATTTTTTCC−10
SSU0984 Cation efflux family proteinPNoNSYes3.02999879 … 1000311 (433)ORFNSNSNSNS
SSU0984 Cation efflux family proteinPNoNSYes3.02999465 … 999906 (442)ORFNSNSCTTTTTCAATAAGTTTTCATC654
SSU0844 Haloacid dehalogenase-like hydrolaseRNo2.65YesNS872555 … 872861 (307)ORFNSNSNSNS
SSU0886 ABC transporter, ATP-binding proteinRYesNSYes2919731 … 920028 (298)ORFAAGGAAAATCTTACAC865GATTTTCCTTCATTCTTTCAT874
SSU1093 LipoproteinRYesNSYesNS1110450 … 1110699 (250)PromNSNSNSNS
SSU1350/1351 PhosphoesteraseRNoNSYesNS1377350 … 1377649 (300)ORFNSNSTTTTTTCAATAATTTTACTGT32
SSU1578 Gamma-glutamyl hydrolaseRNo2.61YesNS1584517 … 1584727 (211)PromNSNSCCCTTTCTTTCATTTTTTCAA−30
SSU1760ssnASurface-anchored DNA nucleaseRYesNSNo−2.451782332 … 1782752 (421)PromGAGAGAACGGTTTCTA8NSNS
SSU1804 ssDNA-binding proteinRNoNSYesNS1830024 … 1830447 (423)ORFNSNSTTTTTTCAATGATTTTCTGAT1033
SSU0600 Low temperature requirement A proteinSNoNSYesNS626837 … 627278 (442)ORFNSNSGATTTTCTGGATATTTTCACC225
SSU1922 Membrane proteinSNoNSYesNS1956800 … 1956999 (200)PromNSNSNSNS
SSU0421 BipA family GTPaseTNo−3.12YesNS453717 … 453978 (262)PromNSNSNSNS
SSU0421 BipA family GTPaseTYes−3.12YesNS455450 … 455799 (350)ORFNSNSNSNS
SSU0688* Crp family regulatory proteinTYes6.4No2.48716794 … 717064 (271)PromATGATAACGCTTTCTT−48NSNS
SSU1878relAGTP pyrophosphokinaseTKYesNSYesNS1908279 … 1908768 (490)PromNSNSNSNS
SSU0093 Preprotein translocase SecY subunitUNoNSYes2.9782542 … 83337 (796)ORFNSNSNSNS
SSU0911* ABC transporter ATP-binding membrane proteinVNoNSYes3.07941411 … 941864 (454)ORFNSNSCATTTTCCTAACCATTTTCAC452
SSU1589 Type I restriction-modification system S proteinVNoNSYesNS1598837 … 1599188 (352)PromNSNSNSNS
SSU1589 Type I restriction-modification system S proteinVYesNSYesNS1598003 … 1598708 (706)ORFNSNSNSNS
SSU0316 Hypothetical proteinNoNSYesNS336850 … 336949 (100)ORFNSNSNSNS
SSU0556 Hypothetical proteinNoNSYes−2.07583784 … 584153 (370)ORFNSNSATTTTTCAAGCCTATTTCAGT240
SSU0592 Putative exported proteinYes−2.06YesNS618565 … 618763 (198)PromNSNSNSNS
SSU0655 Hypothetical protein SSU0655NoNSYesNS680960 … 681338 (379)ORFNSNSNSNS
SSU0718 Putative exported proteinNoNSYesNS751800 … 752099 (300)ORFNSNSGATTTTCAGAAAGATGTTCAA320
SSU0795 Conserved hypothetical proteinNoNSYes2.88827250 … 827499 (250)ORFNSNSNSNS
SSU0821 Putative membrane proteinNoNSYes2.97851733 … 851994 (262)ORFNSNSTCTTTTCATGCACCTTTTCAT135
SSU0888 Putative membrane proteinYesNSYesNS922350 … 922799 (450)ORFNSNSTCTTTTCAAAATTATTTCTCT663
SSU0974 Putative membrane proteinYes−2.55YesNS992210 … 992486 (227)PromNSNSNSNS
SSU0979 Putative exported proteinNoNSYesNS996050 … 996699 (650)ORFNSNSNSNS
SSU1024 Putative membrane proteinNoNSYesNS1039200 … 1039599 (400)ORFNSNSGTTTTTCCTGAAAATTCCGTA516
SSU1110 Putative membrane proteinYesNSYesNS1126280 … 1126294 (415)PromNSNSNSNS
SSU1231slySuilysinNoNSYesNS1261500 … 1262099 (600)PromNSNSGTTTTTCCAACAATTTCCGGA−164
SSU1250 Putative pseudogeneYesNSYes−2.061281800 … 1282109 (310)PromNSNSNSNS
SSU1303 Putative lipoproteinNoNSYes−2.851334272 … 1334551 (280)ORFNSNSTTTTTTCAAGAAAATTTACAA444
SSU1689 Hypothetical protein (pseudogene)NoNSYes2.451703550 … 1703649 (100)ORFNSNSNSNS
SSU1837 Hypothetical protein SSU1837NoNSYesNS1861400 … 1861899 (500)ORFNSNSNSNS

The genome-wide distribution of CBD in early exp and early stat grown S. suis is shown by the circular plot in Fig. 5A. As illustrated by the Venn diagram (Fig. 5B) the ChIP-seq data revealed that the relative distribution of CBD in exp a compared to stat growth phase was similar to that observed in the microarray experiment. Thus, we identified 58 significantly enriched CBD during the exp growth phase and 101 enriched CBD in stat growth phase. Forty-four CBD were detected in both growth phases. Fourteen CBD and 57 CBD were found only at exp or at stat growth respectively.

Figure 5.

Characteristics of the bound DNA loci belonging to the CcpA regulon of S. suis.

A. Circular plot indicating the enriched genomic loci as obtained from the ChIP-seq experiment (scale in base pairs). Concentric circular rings from the outside to the inside. Annotated genes coloured according to predicted function in S. suis P1/7 are shown on a pair of circles, representing both coding strands: dark blue, pathogenicity/adaptation; black, energy metabolism; red, information transfer; dark green, surface structures; cyan, degradation of large molecules; purple, degradation of small molecules; yellow, central/intermediary metabolism; pale blue, regulators; pale green, unknown; orange, conserved hypothetical; brown, pseudogenes; pink, phage/IS elements; next circular ring: G+C% content in black, above average; grey, below average; green circular ring: enriched peaks after ChIP with CcpA antiserum and chromatin of exponential grown S. suis; orange circular ring: enriched peaks after ChIP with CcpA antiserum and chromatin of stationary grown S. suis.

B. Venn diagram illustration of significantly enriched genome regions during early exp and early stat growth of S. suis as deduced from ChIP-seq.

C. ChIP-qPCR of a selected subset of genes bound by CcpA during growth of S. suis. ChIP assays were performed for S. suis strain 10 grown to time points as indicated in Fig. S5. Chromatin was precipitated using an α-CcpA antiserum and preimmune serum respectively. The relative fold changes of DNA enrichment after ChIP-qPCR were calculated from at least two independent experiments and normalized to non-bound gdh promoter DNA amounts. The dotted line indicates the cut-off value to consider DNA loci as ChIP-positive or ChIP-negative respectively.

D. Consensus motifs generated from MEME analysis on all significant enriched peaks listed in Table 1. The height of each letter represents the occurrence frequency at each location. The upper sequence represents the identified S. suiscre motif with a high similarity to the common pseudo-palindromic cre consensus motif and the lower sequence illustrates the newly identified CcpA-binding motif cre2.

CcpA binding was found for either promoter–operator regions or the coding regions of genes. Only two gene regions, dexB and SSU0421, were bound in both. Overall, during stat growth the ratio of CBD corresponding to promoter regions and CBD corresponding to coding regions was lower (promoter: 49.51%, coding region: 50.49%), compared to the CBD of exp grown S. suis (promoter: 74.14%, coding region: 25.86%).

Relating the CBD to the gene expression results from the microarrays in strain 10ΔccpA we found that among the 115 CcpA bound gene regions 64 genes/operons (55.6%) were differentially expressed in either one or both growth phases. Twenty-three genes (39.7%) of the 58 enriched CBD during the exp growth and 36 genes (35.6%) of the 101 CBD enriched in stat growth phase were differentially expressed in strain 10ΔccpA respectively. Furthermore, 27 CBD were associated with genes that were differentially expressed in strain 10ΔccpA during stat growth solely. Within this group we found, e.g. the glycolytic genes fba and tpiA, as well as two putative surface-located nucleases (SSU1355 and SSU1760) to be positively regulated by CcpA in this growth phase. In contrast genes encoding for a putative amino acid transporter (SSU1194), a cation efflux protein (SSU0984), or a putative transcriptional regulator (SSU0735) were repressed.

Interestingly, among the top 25 genes/operons with highest fold changes in the microarrays (Table S5) which seemed to be repressed by CcpA mediated CCR, the respective regulatory elements of 8 genes (SSU1915, SSU0688, glgCAB, glgP, SSU0395, pgmA, glpK, and SSU0681) were bound by CcpA during exp growth. Six gene clusters, including arcABC and manLMN (see also Fig. S4B), turned out not to be directly regulated by CcpA even though their promoter sequences contain predicted high-scored pseudo-palindromic cre sites.

When relating CcpA binding to CCA, CcpA binding was found to be associated with 7 of the top 25 genes/operons suggested to be activated by CcpA. For instance, CcpA binding to regulatory regions of gla, SSU0421, SSU0892, SSU1179, SSU1364, SSU1830, and the first gene (SSU0515) of capsule synthesis cluster was found indicating CcpA mediated CCA for the regulation of these genes.

To analyse more precisely the binding of CcpA during growth we performed CcpA-ChIP with S. suis cultures of early and late exp and early and late stat growth phases. ChIP-qPCR analyses of the precipitated DNA (Fig. 5C) revealed CcpA binding in all growth phases to the loci of some of the selected genes (glgC, eno, SSU0395), whereas for other genes (cps2A, acnA) binding of CcpA was restricted to the early exp growth phase. For the malX locus alternative CcpA binding was found as we observed prominent binding for the early exp and late stat growth but not for the early stat growth phase. These results show that the DNA binding capability of CcpA seems to be continuously present during growth of S. suis; however, CcpA binding depends on the respective gene locus.

Overall these analyses revealed that only approximately 10% of the differentially expressed genes were directly controlled by CcpA. Furthermore, differential binding patterns of CcpA during growth support our hypothesis that in vivo the regulatory capacity of CcpA most likely depends on cofactors.

Characterization of CcpA bound DNA sequences

The above data showed direct binding of CcpA to regulatory sequences for only a minor group of CCR and CCA subjected genes. Furthermore, ChIP-seq revealed that the presence of a B. subtilis homologous pseudo-palindromic cre consensus motif in certain genes does not necessarily coincide with direct binding of CcpA in vivo. This prompted us to analyse all 115 ChIP-positive DNA regions for a conserved common CcpA binding consensus motif. Interestingly, using the MEME suite (Bailey et al., 2006) we were able to define two conserved motifs with highest scores (Fig. 5D). The first was the pseudo-palindromic motif WWG3AAARC8G9YTTTC14WW (W is A or T, R is A or G, Y is C or T) with homologies to the common cre consensus sequence of B. subtilis. The second consensus motif TTTTYHWDHHWWTTTY (H is A or C or T, D is A or G or T) was a symmetric TA-rich sequence characterized by two dominant TTTTY boxes that has, to the best of our knowledge, not been described for CcpA so far.

Next we applied MEME to ChIP-positive DNA regions enriched at both growth phases. In addition MEME was run separately with the CcpA bound DNA regions of the ChIP-positive genes of early exp and early stat growth. As shown in Table S6, comparison of the 58 sequences of the ChIP-positive genes of exp growth revealed a motif (WGAAARMRYTTYCA, M is A or C) with high homology to the common pseudo-palindromic cre consensus motifs. In addition a highly homologous sequence motif (ARAGAWAACGBTTDC, B is G or C or T) was identified when specifying the search on ChIP-positive DNA regions bound during exp growth only. In contrast, a conserved sequence (TTTTCWHHDDWTTTTY) strongly resembling the symmetric TA-rich motif was identified for the 101 ChIP-positive genes of the stat growth phase and for the loci enriched during stat growth only (TTTTCWDDHHWWTTYY). Interestingly, when applying the search criteria to enriched promoter–operator regions, the MEME search resulted in the discovery of the pseudo-palindromic core sequence (GWAARYGYTTTCW). The symmetric motif, however, was predicted when MEME was applied for the enriched regions of open reading frames only or genes with no differential expression levels as input data sets.

Following, both motifs were run against the annotated S. suis genome sequence using the FIMO software. Applying stringent search criteria (P < 1 × 10−5), FIMO detected the pseudo-palindromic motif 67 times within the S. suis genome. Twenty-seven (40.29%) of these predicted loci were already identified by ChIP-seq. Interestingly, all except one loci harboured the identified pseudo-palindromic cre motif within the non-coding promoter–operator region. Furthermore, all CBD of CcpA repressed as well as of CcpA activated genes, were significantly enriched during exp growth, indicating a high-affinity binding site bound by CcpA at this time point. For the second motif (TTTTCWDDHHWWTTYY), which we designated cre2, we obtained 188 identified putative binding sites with a much broader distribution throughout the genome of S. suis. Among these, 45 regions (23.93%) were identified as CBD, as shown by the ChIP-seq approach (Table 1). In contrast to the pseudo-palindromic motif, the symmetric motif was present in 26 cases (57.8%), and thus to a higher extent located in the open reading frames of the respective genes. This observation was exemplarily confirmed by ChIP-PCR for three genes (SSU0984, SSU1024, SSU1194) that harbour the cre2 motif within their coding regions (Fig. S4B). Interestingly, for six genes both binding motifs were found in the CBD, but in none of these one motif overlapped the other. Moreover, 50 CBD could not be related to differential transcript levels of their genes in strain 10ΔccpA. In this subset, instead of the pseudo-palindromic motif, MEME-based analysis revealed the symmetric motif TTTTYMWDDMWWTTTY as a conserved sequence with a higher score than the pseudo-palindromic (Table S6). By reducing the MEME search to differentially expressed genes with CBD regions harbouring a predicted pseudo-palindromic cre or the novel cre2 motif we found conserved sequences strongly resembling the pseudo-palindromic cre consensus motif and the symmetric TA-rich cre2 motif respectively (Table S6).

Taken together, the motif analyses revealed a new symmetric CcpA-binding motif, cre2, which underlines existence of yet unknown regulator function of CcpA.

Characterization of the in vivo relevance of CcpA binding to the cre2 motif

To further elucidate the binding of CcpA to the cre2 motif we performed EMSA with a double-stranded oligonucleotide harbouring the alternative cre2 site (TTTTCAAAAAATTTTC, Table S1), as well as with the respective cre2 MUT oligonucleotide with a point mutated motif (T2T3 to GG and T13T14 to GG). As shown in Fig. 6A, the EMSA of cell lysates obtained from S. suis grown to the early exp and the early stat growth phase clearly demonstrated binding activity of CcpA to the radioactively labelled cre2 oligonucleotide (lane 1 and 9 respectively), which could be competed with a 200-fold excess of the respective unlabelled oligonucleotide (lane 2 and 10) but not by the respective unlabelled mutated oligonucleotide (lane 3 and 11). Addition of an α-CcpA antiserum to the binding reaction before native gel electrophoresis prevented protein DNA complex formation indicating CcpA as the major binding protein (lane 8 and 16). Next, we were interested if the CcpA binding affinity to the cre2 motif was restricted to S. suis. For this, we tested respective lysates from a selected human pathogenic S. agalactiae and S. pyogenes strain for binding to the glgC and to the cre2 oligonucleotide representing a pseudo-palindromic cre motif and the symmetric motif respectively. As shown in Fig. 6B, protein complexes containing the CcpA homologues of S. agalactiae and S. pyogenes, respectively, were able to bind to both the pseudo-palindromic cre motif and the cre2 motif under our conditions, though with different intensities. These data indicate that the newly identified cre2 motif is bound by CcpA of S. suis with high affinity in vitro and that this consensus sequence is relevant for CcpA binding also in other streptococcal species.

Figure 6.

Functional analysis of the cre2 motif.

A and B. EMSA of native CcpA binding in bacterial lysates. EMSA was carried out with the radiolabelled cre2 or glgC oligonucleotides and 10 μg of bacterial cell lysate. Competition experiments in the presence of a 200-fold molar excess of non-labelled unmodified cre2, cre2MUT, glgC or arcA oligonucleotides. Immunoshift assays were performed with α-CcpA antiserum or preimmune serum respectively. (A) EMSA of native CcpA from S. suis wild-type (WT) or strain 10ΔccpA (ΔC) lysates with a radiolabelled oligonucleotide representing the newly identified CcpA-binding motif cre2. (B) EMSA of native CcpA from S. suis, S. agalactiae or S. pyogenes lysates with radiolabelled glgC or cre2 oligonucleotides representing a common pseudo-palindromic cre motif and the newly identified CcpA-binding motif cre2 respectively.

C. GFP reporter assay with S. suis wild-type strain 10 and strain 10ΔccpA carrying the GFP under control of the eno promoter region with or without point mutated cre2 motifs respectively (10::eno_cre2MUT-gfp; 10ΔccpA::eno_cre2MUT-gfp or 10::eno-gfp; 10ΔccpA::eno-gfp). Bars represent the relative fluorescence units (RFU) after normalization to the values obtained for strain 10 carrying the promoterless gfp construct (10::gfp). Means and standard deviation from three independent experiments performed in triplicates are shown.

D. GFP reporter assay with S. suis wild-type strain 10 and strain 10ΔccpA carrying the GFP under control of the glgC promoter region with point mutation within the cre motif (10::glgC_creMUT-gfp; 10ΔccpA::glgC_creMUT), the cre2 motif (10::glgC_cre2MUT-gfp; 10ΔccpA::glgC_cre2MUT-gfp), or both motifs (10::glgC_cre/cre2MUT-gfp; 10ΔccpA::glgC_cre/cre2MUT-gfp). Bars represent the relative fluorescence units (RFU) after normalization to the values obtained for strain 10 carrying the promoterless gfp construct (10::gfp). Means and standard deviation from three independent experiments performed in triplicates are shown.

Following we performed GFP reporter studies with the cre2 motif carrying the eno promoter region as a CcpA activated gene and the cre and cre2 bearing promoter of the glgC gene, as a CcpA repressed gene to proof the in vivo relevance of the cre2 motif. For this, we cloned the eno promoter harbouring the cre2 motif, TTTTCTTGAAAAATCC (Table S1), and the promoter with a point mutated cre2 motif, TGGTCTTGAAAAGGCC, into the gfp reporter plasmid and transformed it into S. suis wild-type strain 10 and strain 10ΔccpA. Furthermore, we point mutated the cre2 motif, TTTTCATTTTTCTTTTC, of the glgC promoter to TGGTCATTTTTCTGGTC in the reporter plasmids used for the experiment depicted in Fig. 4A. As shown by the GFP reporter assay in Fig. 6C, GFP expression of strain 10::eno_cre2MUT-gfp was significantly affected by mutating the cre2 motif. This effect was also seen in strain 10ΔccpA::eno-gfp, indicating a general, CcpA independent role of cre2 for eno expression. However, GFP expression of strain 10ΔccpA::eno-gfp was significantly lower than in strain 10::eno-gfp which suggests a complementary contribution of CcpA to eno expression via cre2. The analyses of the glgC promoter revealed that the mutation of the cre2 motif in the presence of an intact cre site (10::glgC_cre2MUT-gfp) did not relieve the promoter activity from CcpA repression seen in 10::glgC_creMUT-gfp. This indicated that in the context of the glgC promoter the pseudo-palindromic cre motif dominates the cre2 motif. On the other hand, concomitant mutation of the cre and the cre2 motif in the glgC promoter (10::glgC_cre/cre2MUT-gfp) resulted in a significant lower induction of GFP expression compared to 10::glgC_creMUT-gfp. This result indicates that cre2 is necessary for the activation of glgC expression. Reporter analyses performed with strain 10ΔccpA confirmed that CcpA represses glgC expression. The expression level of strain 10ΔccpA::glgC-gfp, however, was significantly higher than that of the mutants in which either the cre, cre2 or both were inactivated (Fig. 6D). This indicates that both cre sites contribute to glgC promoter activation in the absence of CcpA.

Discussion

Metabolic adaptation to different host environments is a critical step for pathogens during establishment of infection. Similar to non-pathogenic bacteria, the capacity to adapt to different carbon sources such as carbohydrates seems to be of major importance to ensure survival and growth in the host. Therefore, these processes are tightly regulated in a hierarchical order by feedback mechanisms such as carbon catabolite repression (CCR) and carbon catabolite activation (CCA) (Seshasayee et al., 2006). The catabolite control protein A (CcpA) is the major transcriptional mediator of CCR and CCA in Gram-positive bacteria. Our previous results indicated a considerable impact of CcpA in the exp growth phase when S. suis has access to glucose (Willenborg et al., 2011). Furthermore, recent cDNA microarray experiments performed with different bacterial species suggest a regulatory role of CcpA in carbohydrate exhausted conditions (Lulko et al., 2007; Carvalho et al., 2011). Therefore, and since CcpA is constitutively expressed in S. suis (Willenborg et al., 2011), in the present study we attempted to dissect the role of S. suis CcpA during growth in more detail.

We observed globally influenced gene expression during early stat growth phase in strain 10ΔccpA suggesting a considerable relevance for CcpA at this growth phase. Expression of 86 genes was also affected at early exp growth. The majority of these genes could be assigned to carbohydrate transport and metabolism which was in correlation with previous findings for the exp growth phase (Shelburne et al., 2008; Carvalho et al., 2011). Hence, this group indicates that CcpA contributes to gene repression and activation in both growth phases of S. suis. In contrast, COG clustering of the 303 genes affected during stat growth phase only did not reveal any cumulative assignment into specific functional groups. In addition, the levels of fold changes in gene expression of the majority of genes in stat phase were not as pronounced as in exp growth phase of S. suis. These findings suggest diverse regulatory roles for CcpA during both growth phases.

ChIP-seq analysis and subsequent in silico analyses of the CcpA bound DNA sequences revealed two putative CcpA consensus sequences in S. suis. The first cre consensus sequence shares high homologies with the B. subtilis pseudo-palindromic cre motif WTGNNARCGNWWWCAW including the highly conserved nucleotides G3, C8, G9, and C14 (Miwa et al., 2000; Fujita, 2009). This is in line with a recently published analysis of cre consensus motifs identified by CcpA derived ChIP-on-chip experiments with B. subtilis exposed to a glucose pulse (Buescher et al., 2012). On the other hand, we identified a second not yet described CcpA-binding motif (TTTTYHWDHHWWTTTY), which we named cre2. Specific binding of CcpA to this motif was confirmed by EMSA, also for homologous proteins of other streptococci. Our EMSA experiments revealed that at least the core thymidine nucleotides at position T2T3 and T13T14 mediate CcpA binding. In addition, by EMSA we found that cre and cre2 were bound by CcpA in lysates of S. suis from both growth phases. Furthermore, the ability of the cre oligonucleotide to compete CcpA binding to the cre2 oligonucleotide and the similar migration of the cre and cre2 oligonucleotide CcpA complexes in our EMSA indicate similar sterical complex formation. This suggests that the CcpA DNA-binding domain is able to bind both cre sites. Furthermore, protein binding to both motifs was shown in lysates of other streptococcal species. These data indicate that cre2 may also be of functional relevance in other streptococci.

To gain insights into the role of the two binding sites we correlated ChIP-seq data with the microarray data. This revealed that during early exp as well as early stat growth only approximately 10% of the differentially expressed genes were directly controlled by CcpA, which is a lower percentage of genes directly regulated by CcpA than described for B. subtilis or Clostridium difficile (Antunes et al., 2012; Buescher et al., 2012).

When focusing on CBD of exp growth phase, our ChIP-seq data revealed that the CBD of many of these genes harbouring the pseudo-palindromic cre motif was located within the promoter regions. Frequently binding correlated with affected expression of these genes in strain 10ΔccpA, which is consistent with previous reports for the cre localization of genes of CcpA mediated CCR or CCA (Fujita, 2009).

In contrast the cre2 motif was predominantly found in CBD of the stat growth phase. Furthermore cre2 was located most often within open reading frames. Interestingly, expression of the genes harbouring the cre2 motif seemed to be less affected in strain 10ΔccpA or was not influenced as deduced from the microarray data. This was in striking contrast to the pseudo-palindromic motif which appeared to be stronger associated with a stringent gene regulation mediated by CcpA. This indicates a distinct role of CcpA for the regulation of genes controlled by the cre2 motif. Among the genes subjected to CCR or CCA which were directly controlled by CcpA, CcpA binding to its recognition sites often did not correlate with repression or activation of respective gene expression. Thus, we observed sustained binding of CcpA during growth to strongly repressed genes harbouring the pseudo-palindromic cre motif like glgC and yvyD (SSU0395, Fig. 5C). Although CcpA is constitutively expressed in S. suis, this finding indicates that for regulating gene expression at least in some cases, CcpA requires one or more cofactors to exert gene regulation. Nevertheless, it might also possible that its regulatory capacity depends on other yet unknown secondary modifications of CcpA itself.

The binding and regulatory activity of CcpA has been shown to be dependent on the serine phosphorylated co-repressor HPr [HPr(Ser)∼P] which is regulated by the HPr kinase/phosphorylase (HPrK/P). Activity of HPrK/P depends on the cytosolic level of fructose-1,6-bisphosphate (FBP) (Fujita, 2009). Interaction of HPr(Ser)∼P with CcpA enables CcpA binding to the promoter or operator region (Deutscher et al., 2006). This seems to be the principal mode of action for CcpA. Our Western blot analyses revealed differential banding patterns of HPr proteins during growth of S. suis, which was also obvious for the heat stable CcpA interacting HPr(Ser)∼P form. Correspondingly, we were able to detect decreasing amounts of CcpA-bound HPr during growth of S. suis. These results suggested that CcpA–HPr(Ser)∼P complex exerting CCR and CCA is present mainly at early exp growth phase which is in agreement with the data from other bacteria (Deutscher et al., 1995; Fujita, 2009). Nevertheless, CcpA of S. suis exhibits constitutive DNA binding activity to regulatory elements of certain genes during growth. Since decreasing amounts of the CcpA–HPr(Ser)∼P complex during growth did not correlate with binding to CBD in vivo as shown in our ChIP-seq and ChIP-qPCR experiments we assume that HPr(Ser)∼P by interacting with CcpA does enhance binding activity of CcpA to the pseudo-palindromic cre motif. Nevertheless, in our study we were not able to identify genes that were bound by both CcpA and HPr by ChIP experiments conducted with a HPr antiserum which might be due to the low amounts of the CcpA–HPr(Ser)∼P complex relative to the unbound complex components. In addition, and similar to studies of other groups working with streptococcal HPr (Zeng and Burne, 2010), we failed to generate a HPr mutant of S. suis to prove our hypothesis. Noteworthy, cofactors such as metabolites can influence the regulatory activity of transcriptional regulators without affecting DNA binding (Chaix et al., 2010).

Based on the CcpA binding and expression levels of the cre2 associated genes one might speculate that CcpA binds to and regulates through the cre2 site independently of HPr(Ser)∼P. Yet a regulatory function of HPr(Ser)∼P bound CcpA associated with cre2 cannot be excluded. For the HPr(Ser)∼P independent cre2 mediated effects CcpA itself most likely needs to be modified or might interact with another unidentified cofactor. Thus, a HPr homologue (SSU1039, ptsH) has been found in S. suis (Dubreuil et al., 1996), but its putative association with CcpA has not been proven so far.

Both reporter analyses to functionally characterize the CcpA binding to the cre2 motif point on a regulatory role in vivo. Thus, our reporter studies of the eno promoter clearly indicate that the cre2 binding site is necessary for eno promoter activation. CcpA seems to contribute only complementary, as the promoter activity was significantly reduced in strain 10ΔccpA::eno-gfp but not to the level of strain 10::eno_creMUT-gfp. The glgC promoter studies in the wild-type strain demonstrate that the cre2 motif, which is clearly distinct from and located upstream the cre motif in the glgC promoter, contributed to glgC promoter activation as the double mutation of cre and cre2 resulted in a partial relief of promoter repression compared to the single cre mutation. Furthermore, as the single mutation of cre2 did not affect promoter repression one can conclude that CcpA binding to the pseudo-palindromic cre dominates cre2 mediated activation. Most probably, cre2 mediated promoter activation is regulated by binding of CcpA in a HPr independent conformation. Nevertheless, we cannot exclude the contribution of other regulators, since we were not able to dissect the precise role of CcpA due to the fact that both cre-binding motifs contributed also to promoter activation in strain 10ΔccpA. The results from strain 10ΔccpA, however, suggest a redundant binding of CcpA and other transcriptional regulators to cre and cre2 sites.

Overall the in vivo relevance of CcpA binding to the cre2 motif seems to be supporting rather than essential for gene expression when compared to the cre motif, which might also explain the moderate fold changes of the genes with cre2 sites in strain 10ΔccpA. Nevertheless the precise functional relevance of CcpA binding to the cre2 site remains to be resolved in more detail in future studies.

Approximately 50 of the ChIP-seq identified CcpA bound DNA loci were found to be not transcriptionally affected by a ccpA deletion suggesting that CcpA expresses constitutive binding activity in vivo. However, the sole binding of CcpA may not necessarily be associated with differential regulation. In agreement with this our ChIP-qPCR experiments from different growth phases showed a general CcpA binding capability in vivo, whereas DNA binding of CcpA is varying depending on the respective regulatory element. This further indicates that the regulatory capacity and binding activity of CcpA most likely depends on cofactors as well as on the regulatory composition of the respective element.

Our data also show that the majority of genes with differential transcript levels in strain 10ΔccpA were not directly regulated by CcpA, although several genes harboured putative cre sites in their regulatory regions. Interesting examples are arcA, manL, or SSU1310 (Fig. S4B). This could be due to the fact that either CcpA does not bind the pseudo-palindromic cre site within the promoter context, or another yet unknown regulator competes with CcpA for the binding site. In case of arcA, we provided experimental evidence that its cre site compared to that of the glgC promoter is not relevant for mediating direct CcpA regulation in vivo, despite being recognized by CcpA in vitro. Therefore, the context of the regulatory elements of the endogenous promoter seems to be of major importance for binding of CcpA in vivo. This also indicates that ChIP is a powerful tool for careful interpretation of in vitro and in silico data in studying bacterial transcription factors in general.

Indirect regulatory effects of CcpA may be considered due to its involvement in the expression of the predicted transcriptional regulators (e.g. SSU0688 or SSU0735) and the predicted sigma modulation factor SSU0395 (Drzewiecki et al., 1998; Tagami et al., 2012). In addition, a putative LicT like antiterminator encoded by SSU1310 was markedly lower expressed in the ccpA-deficient strain during exp growth and may contribute to the indirect effects observed in our study (Kruger et al., 1996; Deutscher et al., 2006).

ChIP-seq allowed us to define the CcpA regulon and enabled us to further dissect the role of CcpA as one key regulator of central carbon metabolism and virulence gene expression of S. suis (Fig. 7). S. suis contains a particularly large number of putative sugar uptake systems (14 PTSs and 6 ABC transporter) allowing the metabolization of a wide variety of carbohydrates encountered in its ecological niches. Only two of these, the lactose/tagatose PTS or maltodextrin ABC-transporter, were directly controlled by CcpA-dependent CCR, indicating a rather narrow direct impact of CcpA for the acquisition of alternative carbohydrates. Nevertheless, as depicted in the schematic model in Fig. 7 CcpA directly influences carbohydrate conversions as a regulator of the glycogenic glgCAB operon, the Leloir pathway, or the amylopullulanase (apuA) involved maltodextrin metabolism via a maltodextrin phosphorylase (glgP). In line with our ChIP-seq data a previous study on Transposon-seq mutagenesis in Streptococcus pneumoniae concluded a genetic interaction of CcpA with glgC, malX, yvyD (SSU0395), but not arcA and manL (van Opijnen and Camilli, 2010).

Figure 7.

Schematic model representing the CcpA regulon of S. suis including simplified metabolic routes. Schematic overview of the S. suis CcpA regulon as derived from gene expression and ChIP-seq experiments. Direct regulation by CcpA is depicted by green or orange lines, reflecting confirmed influence during exponential and stationary growth respectively. The grey arrows indicate direct CcpA binding to the promoter region without any apparent regulation. Arrows indicate direct activation and t-bar arrows direct repression by CcpA. Coloured ovals indicate genes (asterisks indicate first genes of an operon) grouped in functional classes as following: purple, carbohydrate metabolism/conversion; dark green, glycolysis; light pink, pyruvate metabolism; brown, incomplete tricarboxylic acid (TCA) cycle; light green, transcription factor/sigma cofactor; ocher, surface exposed/secreted proteins; blue, capsule synthesis; red boxes, carbohydrate uptake/transport. Pathway predictions and gene annotations were conducted with the information from the KEGG and NCBI database: CcpA, catabolite control protein A; SSU0911, ABC transporter ATP-binding membrane protein; SSU1364, branched-chain amino acid ABC transporter; ABC MalX, maltodextrin ABC transport (SSU1915–1918); Gal, galactose transporter (probably SSU1709–1707); GlpF, glycerol uptake facilitator (SSU0677); PTS Glc, glucose phosphotransferase system (probably SSU1583–1585); PTS Lac, lactose phosphotransferase system (probably SSU0892–0893); apuA, extracellular amylopullulanase; sly, suilysin; ssnA, surface-anchored DNA nuclease; cps2A, capsule synthesis regulatory protein; glpK, glycerol kinase; glpO, alpha-glycerophosphate oxidase; galK, galactokinase; glgC, glucose-1-phosphate adenylyltransferase; glgP, maltodextrin phosphorylase; SSU1169, β-fructofuranosidase; SSU1230, putative N-acetylmannosamine 6-phosphate 2-epimerase; SSU1929, putative beta-glucosidase (bglA); SSU1368, glucan 1,6-alpha-glucosidase (dexB); pgmA, phosphoglucomutase/phosphomannomutase; pgi, glucose-6-phosphate isomerase; fba, fructose-bisphosphate aldolase; tpiA, triosephosphate isomerase; eno, enolase; ldh, l-lactate dehydrogenase; alsS, acetolactate synthase; pdhA, pyruvate dehydrogenase; pfl, pyruvate formate-lyase activating enzyme (pflC); ackA, acetate kinase; adhE, alcohol dehydrogenase; acnA, aconitate hydratase; SSU0395, sigma 54 modulation protein (yvyD); SSU0688, Crp-family regulatory protein; SSU0735, PyrR bifunctional regulatory protein; Glc-1-P, glucose-1-phosphate; Glc-6-P, glucose-6-phosphate; DHAP, dihydroxyacetone-phosphate.

In addition, CcpA directly controls the expression of genes encoding for key enzymes of the glycolytic pathway. Hence, CcpA negatively regulated the first enzyme of the glycolysis encoded by pgi during exponential growth. In contrast, tpiA, fba, and eno were positively regulated by CcpA during stat growth. Interestingly, the identification of the cre2 motif in the promoter regions of pgi and eno, and the lack of the pseudo-palindromic cre motif for all genes points to a relevance of this motif for CcpA in its role as a molecular mediator for key reactions of the central metabolism of S. suis. This is also indicated by the importance of CcpA in regulating the mixed-acid fermentation and the fragmentary TCA cycle. Thus, CcpA is directly regulating the expression of genes responsible for carbohydrate conversions (pgmA, SSU1169, SSU1230, SSU1368), the pyruvate dehydrogenase complex (pdhABC), genes of mixed acid fermentation (ldh, pfl, alsS, adhE), and incomplete TCA cycle (acnA).

We have previously shown that CcpA contributes to the expression of virulence associated genes during growth and capsule thickness (Willenborg et al., 2011). Our ChIP-seq revealed direct regulation of the gene encoding for the pore forming protein suilysin and the capsule synthesis gene cluster. As the capsule is at the interface between carbohydrate metabolism and virulence this finding is of importance as differential capsule expression is important for the main steps of the pathogenesis of S. suis, comprising colonization, survival in the bloodstream, and invasion into the target tissue (Fittipaldi et al., 2012).

In conclusion, our studies revealed that CcpA contributes substantially to S. suis metabolism and virulence, by both direct and indirect regulation of genes. The regulon of CcpA bound genes indicates CcpA as an important regulator of carbohydrate-dependent metabolic switches, and on the other hand also as a regulator for a balanced metabolic flux in the central carbon metabolism. As a regulator of carbohydrate-dependent metabolic switches CcpA most likely requires HPr(Ser)∼P as a cofactor for stringent interaction with the pseudo-palindromic cre motif and regulation of genes leading to pronounced CCR and CCA. Obviously, as a regulator for the central carbon metabolism its action is more moderate. This might be implemented by a less stringent conformation allowing CcpA to regulate gene expression via either the pseudo-palindromic cre motif or the secondary cre2 motif. Overall this study contributes to a better understanding of the role of CcpA within the regulatory network of the carbon metabolism of this important pathogen.

Experimental procedures

Bacterial strains and growth conditions

The virulent serotype 2 wild-type strain 10 (WT), the ccpA-deficient strain 10ΔccpA, and the complemented strain c10ΔccpA have been described earlier (Willenborg et al., 2011). Bacteria were cultured overnight at 37°C on Columbia blood agar base (Difco) containing 6% (vol/vol) sheep blood or horse blood supplemented with erythromycin (2 μg ml−1) or spectinomycin (100 μg ml−1), if necessary. For further experiments bacteria were grown in Todd Hewitt Broth (THB, Becton Dickinson Diagnostics) and harvested at indicated time points depicted in Fig. S5.

DNA microarray analysis

Comparison of gene expression between the wild-type and ccpA-deficient strain at early stat growth was done by microarray analyses as described before (Fulde et al., 2011). Quantitative reverse transcriptase PCR (qRT-PCR) for confirmation of the microarray data was performed as described in Supplementary methods. Heatmap visualization of the microarray data was performed with the Mayday software (Battke et al., 2010). All microarray data have been submitted MIAME complied with ArrayExpress under Accession Nos. E-MEXP-2946 and E-MEXP-3832.

HPr phosphorylation profile

Bacteria were grown in THB medium to early exp and early stat growth phase and harvested. The pellet was resuspended in 10 mM Tris-HCl (pH 8.0) supplemented with 1× protease inhibitor AEBSF (Applichem) and 1× HaltTM phosphatase inhibitor cocktail (Thermo Scientific). The bacterial suspension was treated by ultrasonication in the Branson Sonifier 450 (15 min, 4°C, input setting 8) and the lysates were cleared by centrifugation. When indicated, the bacterial lysates were incubated before electrophoresis with 2 μl of λ-Phosphatase (NEB) in the appropriate buffers at 30°C overnight or heated at 70°C for 10 min to detect dephosphorylated or phospholabile HPr forms respectively. All protein extracts (2 μg) were mixed with one volume of 2× loading buffer [0.125 M Tris-HCl (pH 8.8), 0.004% bromophenol blue, 20% glycerol], separated by native non-denaturing gel electrophoresis and transferred onto a PVDF membrane (Landmann et al., 2012). For the determination of total HPr amounts, protein extracts were mixed with 2× Laemmli buffer, heated at 95°C for 10 min and separated on a 12% polyacrylamide SDS gel. Immunodetection of both phosphorylated and total HPr was performed as described for the co-immunoprecipitation experiments.

GFP reporter studies

Transcriptional fusions of the glgCAB- and arcABC-promoter–operator regions to the green fluorescent protein (gfp) were constructed and GFP reporter experiments performed as described in Supplementary methods.

Electrophoretic mobility shift assay (EMSA)

Double-stranded oligonucleotides were obtained by annealing of complementary single-stranded oligonucleotides. Double-stranded oligonucleotides were radiolabelled with [α-32P]-dCTP (PerkinElmer) and Klenow DNA polymerase (NEB), and afterwards purified with the QIAquick® Nucleotide Removal Kit (Qiagen). For EMSA 200 ng of recombinant CcpA (rCcpA), recombinant suilysin (rSLY, negative control) or 10 μg of bacterial cell lysate were incubated for 30 min at room temperature in a final volume of 40 μl of binding buffer [20 mM Tris-HCl, pH 8.0, 1 mM EDTA, 1.75 mM dithiothreitol (DTT), 0.5μg double-stranded poly(dI-dC), 250 ng bovine serum albumin (NEB), 100 mM KCl, and 4% Ficoll 400] with 20 000 cpm of radiolabelled oligonucleotides. When indicated 200-fold molar of specific or unspecific DNA competitors were included in the reaction mixture. For immunomobility shift assays, 1 μl α-CcpA or α-HPr antiserum was added to the binding reaction followed by additional 15 min of incubation. Reactions were loaded on an 8% non-denaturing 0.5× TBE polyacrylamide gel and separated electrophoretically at 120 volts for 1.5 h at room temperature. Then the gel was dried and exposed to X-ray film at −80°C.

Chromatin immunoprecipitation (ChIP) and ChIP-sequencing (ChIP-seq)

For ChIP analysis the wild-type strain 10, strain 10ΔccpA and strain c10ΔccpA were grown in THB media to early exp and early stat growth phase. The in vivo cross-linking of the cultures was done with 0.15 mM EGS [ethylene glycol bis(succinic acid N-hydroxysuccinimide ester)] for 30 min followed by 1% (v/v) formaldehyde for 8 min, and afterwards quenched by addition of glycine in a final concentration of 0.125 molar at room temperature. ChIP and ChIP-qPCR was performed as described in Supplementary methods. ChIP-Sequencing libraries were prepared from 10 ng of immunoprecipitated DNA with the Illumina ChIP-Seq DNA Sample Prep Kit according to the Illumina instructions. Briefly, DNA was fragmented using the Covaris S2 system, the overhangs resulting from fragmentation were converted into blunt ends, an A-base was added to the 3′ end of the blunt phosphorylated DNA fragments, and adapters were ligated to the ends of the DNA fragments. The Illumina Cluster Station hybridized the fragments onto the flow cell and amplified them for sequencing on the Genome Analyzer IIx. The Genome Analyzer sequenced clustered template DNA using a robust four-colour DNA Sequencing-By-Synthesis (SBS) technology. Fragments were sequenced by 36 bp single end. The fluorescent images were processed to sequences using the Illumina Genome Analyzer Pipeline Analysis software 1.8. The sequence output (36 single end short reads) of the Genome Analyzer IIx was transformed to FastQ format, mapped against the genome sequence of reference strain Streptococcus_suis_P1/7 (NC_012925) with BWA (Li and Durbin, 2010).

Two Illumina runs were performed with CcpA-ChIPs of two biological replicates of cross-linked chromatin from exp and stat grown S. suis. Control immunoprecipitations of either wild-type chromatin (exponential growth) with preimmune serum or 10ΔccpA chromatin (exponential growth) with α-CcpA serum were additionally sequenced. Detection of significant enriched peaks was computed with the CisGenome software (Ji et al., 2008). In a two-sample analysis a conditional binominal model identifies regions in which the ChIP reads are significantly enriched relative to control reads. For further two-sample statistical analysis we used the ChIP-seq data set from the immunoprecipitation of the wild-type chromatin with preimmune serum as the control. In both ChIP-seq data sets (CcpA-IP and control) enriched peaks were detected by scanning the mapped reads with a 50 bp sliding window throughout the genome. Significantly enriched peaks were ranked by a false discovery rate (FDR) and a FDR < 0.05 was applied as the cut-off value. Circular plot of the ChIP-Seq data was done with the DNAPlotter software (Carver et al., 2009). The mapping files (.bam) were transformed using the Mpileup function of the SAMtools package before plotting a DNAPlotter compiled user plot (Li et al., 2009). DNAPlotter settings were set at graph height of 0.5, window size of 10, step size of 5, and track of 0.6. Visualization of raw.bam files of the ChIP-seq data sets was performed with the Artemis software (Rutherford et al., 2000).

All statistically significant enriched peaks were used for motif analysis using the MEME software (Bailey et al., 2006). The search parameters were set to predefined default options. This resulted in the identification of unbiased overrepresented consensus motifs listed in Table S6. If this motif was distributed along the ChIP-peaks we further used FIMO (Find Individual Motif Occurrences) for motif occurrence in the entire set of CcpA bound sequences setting the P value cut-off to 1 × 10−5.

Co-immunoprecipitation

The preparation of cell lysates from cross-linked and native non-cross-linked bacterial cultures was performed as described above and before (Willenborg et al., 2011) respectively. For immunoprecipitation 250 μg of cross-linked or native cell lysates were incubated with 10 μl of α-CcpA serum in a total volume of 1 ml overnight. Protein-A/G precipitated complexes were washed as in the ChIP assays. Elution was performed in SDS sample buffer at 95°C for 10 min, and the samples were separated by 15% SDS-PAGE and subsequently blotted onto a PVDF membrane (Serva). Membranes were incubated with polyclonal antiserum raised against HPr (diluted 1:2000 in 1% BSA) at room temperature for 60 min. Membranes were developed with horseradish peroxidase-conjugated anti-rabbit IgG (Amersham) diluted 1:10 000 in 1% BSA and the SuperSignal West Pico Chemiluminescent Substrate (Pierce) as described by the manufacturer.

Acknowledgements

We gratefully acknowledge Hilde Smith (Central Veterinary Institute, Wageningen University, Lelystad) for providing S. suis strain 10 and Susanne Talay (Helmholtz Centre for Infection Research, Braunschweig) for providing S. pyogenes strain A40. The authors would like to thank Boris Goerke (Georg-August-University, Göttingen) for his help on the HPr phosphorylation profiles, and Tim Carver (Wellcome Trust Sanger Institute, Cambridge) for his help with the DNAPlotter software.

This work was supported by the Deutsche Forschungsgemeinschaft (DFG, Germany) as part of the Priority Programme SPP1316 (Grant Go983-3/1).

Ancillary