Integration of Metabolism and Virulence by Clostridium difficile CodY

Abstract

CodY, a global regulatory protein that monitors the nutrient sufficiency of the environment by responding to the intracellular levels of GTP and the branched-chain amino acids, was previously shown to be a potent repressor of toxin gene expression in Clostridium difficile during growth in rich medium. In the intestinal tract, such derepression of toxin synthesis would lead to destruction of epithelial cells and the liberation of potential nutrients for the bacterium. CodY is likely to play an important role in regulating overall cellular physiology as well. In this study, DNA microarray analysis and affinity purification of CodY-DNA complexes were used to identify and distinguish the direct and indirect effects of CodY on global gene transcription. A codY null mutation resulted in >4-fold overexpression of 146 genes (organized in 82 apparent transcription units) and underexpression of 19 genes. In addition to the toxin genes, genes for amino acid biosynthesis, nutrient transport, fermentation pathways, membrane components, and surface proteins were overexpressed in the codY mutant. Genome-wide analysis identified more than 350 CodY binding regions, many of which are likely to correspond to sites of direct CodY-mediated regulation. About 60% of the CodY-repressed transcription units were associated with binding regions. Several of these genes were confirmed to be direct targets of CodY by gel mobility shift and DNase I footprinting assays.

The ability to sense and adapt to excess or limiting nutrients is a universal trait in the bacterial world. For Clostridium difficile, a Gram-positive, anaerobic, spore-forming, intestinal bacterium, this ability may be essential for its proliferation in the gastrointestinal tract (18). In the hospital environment, C. difficile is now the most common cause of antibiotic-associated colitis, a disease that has increased in incidence and severity in recent years due to the emergence of highly virulent “epidemic” C. difficile strains (19). Antibiotic treatment is thought to induce C. difficile infection (CDI) by disrupting the colonic microflora, thereby allowing C. difficile to colonize (19). Most highly virulent C. difficile strains produce two large protein toxins, A and B, that act by glycosylating members of the host cell's Rho family of small GTPases (16, 17, 52). Recent work of Lyras and colleagues (25) has shown that toxin B, but not toxin A, is essential for virulence in a hamster model of CDI. Some C. difficile strains, including the current epidemic strain type NAP-1/027, also produce a binary ADP-ribosylating toxin, CDT, that is reported to increase colonization (42) but is not known to be essential for pathogenicity (19, 32).

The genes encoding toxins A and B (tcdA and tcdB) lie within a 19.6-kb pathogenicity locus (PaLoc) that also includes the tcdR, tcdE, and tcdC genes (4, 11). In broth cultures of C. difficile, expression of tcdA and tcdB is induced when nutrients become limiting and cells enter stationary phase (8, 14). A clear link between nutrient limitation and toxin gene expression came from the discovery that expression of all the PaLoc genes is repressed by the global transcriptional regulator CodY (7). CodY represses toxin gene expression by binding to the putative promoter region for the tcdR gene (7), whose product is a sigma factor that directs transcription from the tcdA and tcdB gene promoters as well as from its own promoter (30, 31).

CodY proteins, first discovered in Bacillus subtilis and found in many other low-G+C-content Gram-positive bacteria, appear to have in common the ability to repress during rapid growth genes whose products are not needed when nutrients are in excess and to release this repression under nutrient-poor conditions (46). One interpretation of this effect is that CodY proteins help to regulate the synthesis and distribution of pyruvate and 2-oxoglutarate, two key intermediates in central metabolism (47). Functions commonly regulated by CodY in various bacteria include carbon overflow metabolism; the Krebs cycle; synthesis of certain amino acids; uptake and catabolism of amino acids, peptides, and sugars; genetic competence; motility; and sporulation (3, 6, 10, 12, 21, 33, 37). For most of these genes, CodY acts as a transcriptional repressor. Only in the case of the B. subtilis ackA gene has CodY been shown to act as a direct positive regulator (44).

CodY proteins also have species-specific roles, particularly with respect to virulence gene expression, in C. difficile (7), Staphylococcus aureus (27, 28, 37), Streptococcus pneumoniae (12), Streptococcus pyogenes (29), Streptococcus mutans (21), Listeria monocytogenes (3), Bacillus cereus (13) and Bacillus anthracis (51). Although the number of genes whose expression is affected by a codY mutation in any given bacterium is large, in only a few cases has the direct binding of CodY to the regulatory region of a regulated gene been demonstrated. As a result, our knowledge of the extent to which the CodY regulon is defined by the direct action of CodY, as opposed to an effect of CodY on the synthesis or activity of other regulatory proteins, is rather limited.

Structural studies of B. subtilis CodY suggest that the dimeric protein binds to DNA through winged helix-turn-helix motifs located in the C-terminal domain (15, 22). Interaction of B. subtilis CodY with DNA is enhanced in the presence of two different kinds of effectors whose activities account for CodY's response to nutritional conditions. GTP and the branched-chain amino acids (BCAAs) act synergistically to increase the affinity of CodY for DNA (38, 45). C. difficile CodY also responds to GTP and BCAAs in vitro; that is, CodY binding to the tcdR promoter region is enhanced synergistically in response to these effectors (7). A CodY consensus binding motif, AATTTTCWGAAAATT, has been identified upstream of many of the genes regulated by CodY in Lactococcus lactis and B. subtilis (2, 6, 10).

In the work described here, we have identified many genes regulated by C. difficile CodY. Using microarray analysis, we found more than 140 genes that appeared to be directly or indirectly under CodY control. The products of these genes are active in many aspects of metabolism and cell function. Affinity purification was used to pinpoint direct targets of CodY, allowing us to distinguish between transcription units directly and indirectly influenced by CodY. For selected genes, CodY was confirmed to interact directly with their putative promoter regions by electrophoretic mobility shift and DNase I footprinting experiments. Many apparent direct targets of CodY-mediated regulation contain sequences with similarity to the consensus CodY binding motif.

MATERIALS AND METHODS

Strains and growth conditions.

Bacterial strains and plasmids used in this study are described in Table 1. C. difficile strains were grown in TY medium (8), in tryptone-yeast extract (TY) medium supplemented with 1% glucose (TYG) or in brain heart infusion medium (Difco) supplemented with yeast extract (5 g/liter) and 0.1% cysteine (BHIS). Media for growth of C. difficile were supplemented with 10 μg of thiamphenicol ml⁻¹ as needed. C. difficile strains were grown in an atmosphere of 10% H₂, 5% CO₂, and 85% N₂ in an anaerobic chamber (Coy Laboratory Products) at 37°C. Escherichia coli strains were grown at 37°C in L medium (24) supplemented with 50 μg of ampicillin ml⁻¹, as needed.

TABLE 1.

Bacterial strains and plasmids

Strain or plasmid	Characteristics or description	Reference or source
Strains
E. coli
TOP10		Invitrogen
C. difficile
630	Wild-type	53
JIR8094	Ems derivative of strain 630	34
JIR8094(pSD21)	JIR8094 containing pSD21	7
JIR8094::pSD21	codY mutant	7
Plasmids
pJIR1456	E. coli-C. perfringens shuttle vector; Tmr	26
pSD21	Internal region of C. difficile codY gene cloned in pJIR1456; Tmr	7
pCR2.1	E. coli cloning vector; Apr	Invitrogen
pSD30	494-bp DNA fragment containing the CD2344 gene promoter region cloned in pCR2.1; Apr	This study
pSD31	780-bp DNA fragment containing the glgC gene promoter region cloned in pCR2.1; Apr	This study
pSD32	861-bp DNA fragment containing the hisZ gene promoter region cloned in pCR2.1; Apr	This study

DNA preparation.

Genomic DNA (gDNA) was extracted from C. difficile strain 630 that had been grown overnight in 10 ml of BHIS broth. Cells were pelleted and washed in 1 ml of TE buffer (10 mM Tris-HCl [pH 8.0] and 1 mM EDTA) and then resuspended in 400 μl of TE-glucose buffer (50 mM glucose, 25 mM Tris-HCl [pH 8.0], and 10 mM EDTA) containing 25 mg of lysozyme per ml. The suspension was incubated for 2 h at 37°C before addition of N-lauroyl sarcosine to a final concentration of 7%. After addition of 15 μl of RNase A (10 mg/ml), the suspension was incubated for 15 min at 37°C. To this mixture, 10 μl of proteinase K (8 units; Sigma-Aldrich) was added, and the solution was incubated for 30 min at 37°C. TE buffer was used to bring the volume up to 1 ml. The DNA was extracted three times with phenol-chloroform-isoamyl alcohol (25:24:1) and once with chloroform. DNA was precipitated with 3 volumes of isopropanol and 1/10 volume of 3 M sodium acetate (pH 5.2), washed with 70% ethanol, air dried, and resuspended in 100 μl of TE buffer.

RNA preparation.

Overnight cultures of C. difficile grown in TYG medium were diluted to an optical density at 600 nm (OD₆₀₀) of 0.05 in TY medium and incubated anaerobically at 37°C. Samples were removed for RNA isolation at an OD₆₀₀ of 0.4 (exponential phase), diluted with an equal volume of ice-cold acetone-ethanol (1:1), and stored at −80°C. The cell suspensions were thawed and centrifuged at 4°C; the pellets were air dried, washed twice with 500 μl of TE buffer (pH 7.6), and resuspended in 1 ml of buffer RLT from the Qiagen RNeasy kit. Silica glass beads (0.1 mm) were added, and cells were disrupted using a Mini-BeadBeater (BioSpecs Products). Silica beads were removed by centrifugation, and total RNA was isolated from the supernatant fluid using a Qiagen RNeasy kit. DNA contaminating the RNA samples was removed using an Ambion TURBO DNA-free kit for reverse transcription-PCR (RT-PCR) and 5′ rapid amplification of cDNA ends (RACE) analysis. Independent samples of each culture were used to provide DNA templates to verify by PCR the maintenance of the integrated plasmid at the codY locus (oligonucleotides used are listed in Table S1 in the supplemental material). In addition, the characteristic smooth, partly translucent colony morphology of the codY mutant strain was used as an indicator of the retention of the integrated plasmid in the mutant strain culture.

Microarray analysis.

The C. difficile microarray based on the genome sequence of strain 630 was obtained from the Bacterial Microarray Group at St. George's Hospital, University of London (BμG@S), and its design has been described previously (34, 48). The array layout, representing 3,688 of the 3,776 identifiable open reading frames, can be found through the BμG@S website (http://bugs.sgul.ac.uk/A-BUGS-20) and ArrayExpress (accession number A-BUGS-20). A Genisphere 3DNA Array 900 MPX microarray kit was used for cDNA synthesis, labeling, and hybridization with specific protocol modifications as described by O'Connor et al. (34). Ten micrograms of starting RNA was used for the cDNA synthesis reactions. The microarray slides were scanned using a ScanArray 4000 (Packard Instrument Co.), and fluorescence intensities were quantified using ImaGene (BioDiscovery) software. Raw expression data from three arrays, representing three biological replicates, were normalized and averaged using GeneSpring (Agilent Technologies) software. A dye swap was performed with one of the three replicates. A gene was defined as differentially expressed if it exhibited at least a 4-fold change in transcript level in the mutant versus the wild-type strain and if the change was statistically significant by a Student's t test (P < 0.005). All such genes are listed in Table S2 in the supplemental material.

Quantitative reverse transcriptase PCR (qRT-PCR) analysis.

SuperScript II reverse transcriptase (Invitrogen) was used to generate randomly primed cDNA pools from 1 μg of RNA, as instructed by the manufacturer, from each of three strains, JIR8094, JIR8094 (pSD21), and JIR8094::pSD21. To control for chromosomal DNA contamination, mock cDNA synthesis reaction mixtures containing no reverse transcriptase were used as negative controls in subsequent amplifications. cDNA samples were diluted 4-fold and used as templates for quantitative PCR of transcripts from rpoC (primers oMC44/oMC45), ilvC (oMC152/oMC153), CD2344 (oMC154/oMC155), glgC (oMC156/oMC157), hisZ (oMC158/oMC159), speA (oMC160/oMC161), CD1276 (oMC162/oMC163), spoIIE (oMC164/oMC165), CD1476 (oMC166/oMC167), and CD1803 (oMC168/oMC169), using Qiagen SYBR green PCR mix and an MXP3005 thermocycler (Stratagene/Agilent Technologies). The sequences of all primers are listed in Table S1 in the supplemental material. Reactions were performed in a final volume of 25 μl using 4 μl of diluted cDNA, and a 1 μM concentration of each primer. Reactions were performed in triplicate using cDNA extracted from a minimum of two biological replicates. Amplification included 40 cycles of the following steps: 15 s at 95°C, 60 s at 50°C, and 30 s at 72°C. Results were calculated using the comparative cycle threshold method (41), where the amount of target mRNA is normalized relative to an internal control transcript (rpoC), and then that ratio is compared to the ratio in the parental strain (JIR8094).

Affinity purification of CodY-DNA complexes and analysis using the Illumina Genome Analyzer II.

We adapted a method developed by C. D. Majerczyk et al. (27) to screen the C. difficile genome for CodY binding sites. Four samples of 7.5 μg of gDNA were each dissolved in 500 μl of sterile deionized water and sheared by sonication (Branson 250 microtip sonicator) in five 30-s cycles, on ice. Following shearing, gDNA was purified using a Qiagen PCR Purification kit according to the manufacturer's instructions, and resuspended in 30 μl of Qiagen buffer EB. For each sample to be bar-coded, 5 μg of sheared, purified gDNA was blunted using a Quick Blunting kit from New England Biolabs, cleaned with a Qiagen PCR Purification kit, and eluted in 30 μl of buffer EB. A dAMP residue was added (A-tailing) to the 3′ ends of the DNA fragments using dATP and DNA polymerase Klenow fragment (exo-) (New England Biolabs); the fragments were cleaned again with the Qiagen PCR Purification kit and eluted in 30 μl of buffer EB. To generate double-stranded adapters, equal amounts of complementary oligonucleotides (e.g., BC1a and BC1b primers for bar code 1) (see Table S1 in the supplemental material) were mixed, heated to 95°C, and slowly cooled to <40°C. (The b series of oligonucleotides was phosphorylated at the 5′ ends after synthesis). An adapter (0.125 nmol) (see Table S1), specific to each bar code, was then ligated to the sheared, blunted, A-tailed gDNA using a Quick Ligation Kit (New England Biolabs). Each sample was then run on a 2% agarose gel, and DNA corresponding to the 400- to 500-bp size range was recovered from the gel using a Qiagen Gel Extraction kit. Each purified, bar-coded sample was split into 10 aliquots, each of which was subjected to 15 cycles of PCR amplification using primers Olj139 and Olj140 (see Table S1). The PCRs were then cleaned using a Qiagen PCR Purification kit and pooled for each bar code, and the DNA products were quantified.

In a total volume of 250 μl, 5 μg of bar-coded C. difficile strain 630 gDNA was used in a binding reaction with 200 nM purified His₆-tagged C. difficile CodY (7) or His₁₀-tagged B. subtilis aconitase protein (43) (as a negative control) in a binding buffer containing 10 mM (each) isoleucine, leucine, and valine (BCAAs) and 2 mM GTP, 20 mM Tris-Cl, pH 8.0, 50 mM sodium glutamate, 10 mM MgCl₂, 0.05% Nonidet P-40, 5% glycerol, and E. coli tRNA (25 μg/ml). After 25 min at room temperature, the binding reaction mixture was combined with a 0.1-ml packed volume of Talon Co²⁺ Metal Affinity Resin (Clontech) that had been preequilibrated with binding buffer. Following incubation at room temperature for 20 min, the resin was washed six times with 0.25 ml of binding buffer. The resin was then resuspended in 100 μl of 10 mM Tris, pH 8.0, boiled for 5 min, allowed to come to room temperature, incubated with proteinase K (100 μg/ml) at 65°C for 2 h, and then boiled for 5 min. After centrifugation for 1 min at 3,000 × g, 100 μl of the supernatant fluid was transferred to a fresh tube and purified using a PCR purification kit (Qiagen). The sample was then subjected to 10 cycles of PCR amplification using primers Olj139 and Olj40, purified with the PCR purification kit, and quantified. To verify that the pulldown was successful, semiquantitative PCR was performed with primers for the rpoC gene (oMC44/oMC45) and a known CodY target gene, tcdR (oLB51/oLB52). Samples were then subjected to 40-nucleotide (nt), single-ended sequence analysis using the Illumina Genome Analyzer II by the Tufts University Nucleic Acid and Protein Core Facility. Each pulldown experiment was performed in duplicate. CodY-bound and gDNA (control) samples were labeled independently using separate bar codes (1 to 4).

Reference alignment and assembly were performed using MAQ (23) against the C. difficile 630 complete genome reference sequence (NCBI accession NC_009089.1). Using a series of BioPerl scripts, 5′ tag positions were identified and then shifted over by one-half the average DNA fragment length, as determined by an Agilent Bioanalyzer DNA 1000 chip. Tag counts for each position were averaged using a sliding window, creating a coverage map for the entire genome. From this analysis, the CodY/total DNA enrichment ratio at every position was found. In control samples (aconitase-enriched DNA or unselected gDNA), no region was enriched more than 5-fold above background. Regions of enrichment in the CodY sample were considered when the CodY samples showed an average enrichment ratio greater than 5-fold for more than 100 consecutive base pairs. Using a method described by Pepke et al. (35), the false discovery rate at 5-fold enrichment was determined to be 0.01. To focus on the sites most likely to be bona fide targets of CodY, we selected the 350 most highly enriched regions (see Table S3 in the supplemental material), all of which were enriched at least 17-fold in at least one experiment. The calculated z scores for these genes ranged from 13 to 100, indicating that the chance occurrence of this level of enrichment is infinitely small.

The sequences of 293 enriched regions were searched for conserved motifs using the MEME suite, version 4.1.0 (1). The parameters were as follows: 0 to 1 occurrences per fragment on either strand of DNA and a width of 6 to 50 nt. The motif that resulted was found in 280 sequences with an overall P value of 5.2E−68. The sequence with least adherence to the motif had a P value of 3.6E−0.03.

RT-PCR analysis.

Reverse transcription was performed using SuperScript II reverse transcriptase (Invitrogen) with 50 ng of total RNA and 2 pmol of gene-specific primer (see Table S1 in the supplemental material), according to the manufacturer's protocol. Two-microliter samples of the reverse transcription reactions were used as templates for PCR with gene-specific primers (Table S1). PCRs were performed using 27 cycles of denaturation (94°C for 30 s), annealing (50°C for 30 s), and extension (72°C for 1 min). For each gene tested, the amount of RNA used and the number of cycles chosen for the experiment were below saturation for both the reverse transcription and the PCR. As a control for chromosomal DNA contamination, RNA was used directly for PCR amplification without reverse transcription.

Gel mobility shift assays.

DNA fragments containing promoter regions of the ilvC (423 bp) and CD2344 (494 bp) genes were amplified from C. difficile strain 630 chromosomal DNA by PCR using specific primers (OSD107/OSD108 for ilvC; OSD124/OSD125 for CD2344) (see Table S1), agarose gel-purified, and labeled with [γ-³²P]ATP using T4 polynucleotide kinase (Invitrogen) as described by the manufacturer. For amplification of DNA fragments containing the glgC (780 bp) and hisZ (861 bp) promoters, primers OSD127 and OSD128, respectively, were radioactively labeled with [γ-³²P]ATP using T4 polynucleotide kinase (Invitrogen) as described previously (20) and used in conjunction with OSD126 and OSD129, respectively, to amplify the promoter regions. Labeled DNA was mixed with increasing amounts of His₆-tagged CodY protein (7) in 10-μl reaction mixtures that contained 20 mM Tris-Cl (pH 8.0), 50 mM sodium glutamate, 10 mM MgCl₂, 5 mM EDTA, 0.05% (vol/vol) Nonidet P-40 (Igepal; Sigma-Aldrich), 5% (vol/vol) glycerol, and 250 ng of calf thymus DNA. Where indicated, 10 mM (each) isoleucine, leucine, and valine (BCAAs) and 2 mM GTP was included. After incubation for 30 min at room temperature, binding reaction mixtures were loaded on a 12% nondenaturing polyacrylamide gel prepared in Tris-glycine buffer, and electrophoresis was carried out in 35 mM HEPES-43 mM imidazole buffer (pH 7.4), as described previously (45). When 10 mM BCAAs were present in the binding reaction mixture, BCAAs at the same concentration were also added to the electrophoresis buffer. Gels were dried under vacuum and exposed to a phosphorimager screen before detection with a Molecular Dynamics Storm 860 Imager. Gel images were analyzed using ImageQuant, version 1.2, Macintosh software.

DNase I footprinting assays.

DNA fragments containing promoter regions for CD2344 (494 bp), glgC (780 bp), and hisZ (861 bp) were amplified by PCR from C. difficile strain 630 chromosomal DNA using specific primers (OSD124/OSD125, CD2344; OSD126/OSD127, glgC; OSD128/OSD129, hisZ) (see Table S1 in the supplemental material) and then cloned in plasmid pCR2.1 (Invitrogen), yielding pSD30 (CD2344), pSD31 (glgC), and pSD32 (hisZ). Primers OSD124, OSD126, OSD174, and OSD128 were radioactively labeled with [γ-³²P]ATP using T4 polynucleotide kinase (Invitrogen), as described previously (20), and used in conjunction with OSD125, OSD170, OSD126, and OSD169, respectively, to amplify the CD2344, glgC (top strand labeled), glgC (bottom strand labeled), and hisZ promoter regions, respectively, by PCR using C. difficile strain 630 chromosomal DNA as a template. Binding reaction mixtures were prepared as described for the gel mobility shift assays except the reaction volume was 20 μl. All reaction mixtures contained 10 mM (each) BCAAs and 2 mM GTP. After incubation for 30 min at room temperature, 6 mM MgCl₂, 6 mM CaCl₂, and 0.15 or 0.25 U of RQ1 DNase I (Promega) were added to each reaction mixture, and incubation was continued for 1 min at room temperature. The reactions were stopped by the addition of 1 μl of 0.5 M EDTA and transferred to an ice bath. Samples were extracted with phenol-chloroform, and the DNA was ethanol-precipitated and resuspended in 4 μl of sequencing gel loading buffer (40). The samples were heated to 80°C for 5 min and then subjected to electrophoresis in an 8 M urea-6% polyacrylamide gel. Dideoxy sequencing reactions were performed with primers OSD124, OSD126, OSD174, and OSD128 and pSD30, pSD31, or pSD32 as template DNA using a Sequenase kit (US Biochemical) and [α-³⁵S]dATP.

Primer extension.

Total RNA (5 or 10 μg) was subjected to reverse transcription using SuperScript II reverse transcriptase (Invitrogen) and 2 pmol of gene-specific primer (OSD125 for CD2344; OSD170 for glgC) according to the manufacturer's protocol in a 20-μl reaction mixture. Ten microliters of sequencing gel loading buffer was added to the sample, and the samples were heated to 80°C for 5 min and then subjected to electrophoresis in an 8 M urea-6% polyacrylamide gel. Dideoxy sequencing reactions were performed with primers OSD125 and OSD170 using a Sequenase kit (US Biochemical) and [α-³⁵S]dATP.

5′ RACE analysis.

Mapping of the 5′ end of RNA transcripts was performed using 115 ng of total RNA and a 5′ Rapid Amplification of cDNA Ends (RACE) kit (version 2.0; Invitrogen) according to the manufacturer's protocol. First-strand cDNA synthesis was performed with gene-specific primers (OSD179 for hisZ or OSD88 for tcdR). Nested PCR amplification of the first-strand cDNA products was conducted using gene-specific primers (OSD178 for hisZ; OSD181 for tcdR) and the abridged anchor primer supplied with the kit. The nested PCR products were used as templates in sequencing reactions with gene-specific primers (OSD178 for hisZ; OSD181 for tcdR).

Microarray data accession number.

The data discussed in this publication have been deposited in NCBI's Gene Expression Omnibus (9) and are accessible through GEO Series accession number GSE23192 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE23192).

RESULTS AND DISCUSSION

Identification of CodY-regulated genes by microarray analysis.

We used DNA microarray analysis to compare transcript levels in a codY null mutant strain, JIR8094::pSD21, in which the codY gene has been interrupted by integration of an unstable plasmid (7), and a strain that expresses wild-type levels of CodY. The control strain was JIR8094 (pSD21), in which plasmid pSD21 remains primarily extrachromosomal (7). Doing so allowed us to grow both strains in the presence of thiamphenicol. RNA for microarray analysis was isolated from cells grown in TY medium to mid-exponential phase. The cultures were plated before RNA isolation to verify the uniform appearance of colonies with the characteristic smooth, partly translucent appearance of the codY mutant. The cultures were also tested by PCR to verify that the plasmid remained integrated in strain JIR8094::pSD21. Note that we have previously shown that in strain JIR8094::pSD21 the phenotype is due to disruption of the codY gene and not to any polar effect on downstream genes or to any mutation located elsewhere on the chromosome (7).

We identified 146 genes putatively organized in 82 transcription units whose transcripts were at least 4-fold more abundant in the codY mutant than in the strain producing wild-type levels of CodY and for which P values of ≤0.005 were obtained. These genes are listed in full in Table S2 and summarized in Table 2 . In addition, 19 genes in 15 apparent transcription units were underexpressed by more than 4-fold in the codY mutant, with P values of ≤0.005 (Table 2; see also Table S2 in the supplemental material). In agreement with previous results (7), expression of four genes of the PaLoc (tcdR, tcdB, tcdE, and tcdA) was greatly derepressed (79- to 167-fold) in the codY mutant. Most of the other genes of identifiable function that were overexpressed in the codY mutant are involved in metabolism, including biosynthesis of amino acids (isoleucine, valine, leucine, histidine, and arginine); glycogen synthesis; transport of amino acids, oligopeptides, and sugars; proteolysis; synthesis of butyrate from γ-aminobutyrate, succinate, and acetyl-coenzyme A (CoA); ATP synthesis; and sporulation. Most of these functions have been seen to be under CodY control in other bacterial species as well (3, 6, 10, 12, 28, 33). In addition, genes for apparent cell surface (CD0440 and CD1803) and cyclic di-GMP signaling (CD0757 and CD1476) proteins were overexpressed in the codY mutant. In contrast, the genes of the pathway responsible for the conversion of arginine to the polyamines putrescine and spermidine (spe genes) were underexpressed in the codY mutant.

TABLE 2.

Selected genes whose expression was altered in the codY mutant

Gene group, function, and namea	Predicted functionb	Fold changec	Enriched gened
CodY-repressed genes
Amino acid biosynthesis
CD0989-CD0992 (leuACDB)	Leucine biosynthesis	21-99.5	leuA
CD1547-CD1554 (hisZGCBHAF)	Histidine biosynthesis	7.9-15.1	hisZ
CD1565-CD1566 (ilvCB)	Isoleucine-valine biosynthesis	58.1-154.8	ilvC
CD2014 (ilvD)	Isoleucine-valine biosynthesis	31.6
CD2030-CD2034 (argFJC)	Arginine biosynthesis	2.5*-10.6	argF, argC
CD2502	Histidine biosynthesis	34.6	CD2502
CD2726 (glyA)	Glycine biosynthesis	4.7	glyA
CD2828	Aromatic amino acid biosynthesis	18.6	CD2828
Transport
CD0165	Amino acid transporter	31.9	CD0165
CD0324 (cbiM)	Cobalt transporter	5.16
CD0388-CD0390 (bglPAG)	β-Glucoside PTS transporter	14.6-19.2	bglP
CD0449	Na+/H+ antiporter	7.5
CD0853-CD0857 (oppBCADF)	Oligopeptide ABC transporter	12.6	oppB
CD0875-CD0873	ABC transporter	3.1*-4.9	CD0875
CD0878-CD0876	ABC transporter	4.5-15.6	CD0878
CD0879	Carbohydrate kinase	7.5	CD0879
CD1259-CD1260	Branched-chain amino acid transporter	4.9-23.7	CD1259, CD1260
CD2511-CD2508	Sugar transport system	3.6*-10.1	CD2511
CD2512	PTS system, IIa component	4.0
CD2702 (brnQ)	Branched-chain amino acid transporter	6.8	brnQ
CD3615	MerR-family transcriptional regulator	5.5
CD3616	Na+/H+ exchanger	7.0
CD3628	Sugar transporter	6.2	CD3628
Sporulation
CD0770-CD0772 (spoIIAA-AB-AC)	Anti-sigma F factor antagonist	7.9-12.7
CD3490 (spoIIE)	Stage II sporulation protein	17.1
CD3516 (spoVG)	Sporulation protein VG	4.87
Protein degradation
CD1086	Peptidase	57.1
CD2822	Peptidase	89.5	CD2822
CD2862	Dipeptidase	21.6
Fatty acid and membrane synthesis
CD1054-CD1059 (bcd2 etfB2 etfA2 crt2 hbd thlA1)	Butyrate metabolism	21.9-67.1	bcd2
CD1327 (pgsA)	Phosphatidylglycerophosphate synthase	607.7
CD2426-CD2425	Butyrate kinase	4.6-6.6	CD2426
Energy metabolism
CD2960-CD2954 (ntpIKECABD)	V-type sodium ATP synthase	14.4-89.3	ntpC, ntpI
Other metabolism
CD0444	Oxidoreductase	10.4
CD0445-CD0447	d-Ornithine utilization	3.9*-15.4	CD0445
CD0849 (abgB2)	Aminobenzoyl-glutamate utilization	7.7	abgB
CD0882-CD0886 (glgCDAP)	Glycogen biosynthesis	6.1-17.8	glgCD
CD1263	FMN-dependent dehydrogenase	4.1	CD1263
CD1389	Chloromuconate cycloisomerase	160.3
CD1612	Amidohydrolase	18.9
CD2158 (gabT)	4-Aminobutyrate aminotransferase	31.9	CD2158
CD2164 (ldh)	l-Lactate dehydrogenase	6.8	CD2164
CD2344-CD2340	Gamma-aminobutyrate/hydroxybutyrate metabolism	26.7-95.1	CD2344
CD2819	Amino acid racemase	5.6	CD2189
CD2859	d-Aminoacylase	4.3	CD2859
Chromosome replication and cell division
CD1328 (recA)	RecA protein (recombinase A)	211.6
CD1329	Nucleic acid-binding protein	103.8
Virulence
CD0440	Cell surface protein	4.46	CD0440
CD0659-CD0663 (tcdRBEA)	Sigma factor, toxin B, holin, toxin A	51.7-174.8	tcdR, tcdB
CD1803	Cell surface protein	23.79	CD1803
Signaling and regulation
CD0616	MerR-family transcriptional regulator	7.32
CD0618	Transcriptional regulator	53.72
CD0757	c-di-GMP signaling protein	4.49
CD1265	Regulatory protein	4.48
CD1476	c-di-GMP signaling protein	4.12	CD1476
CodY-activated genes
CD0106 (cwlD)	Germination-specific N-acetylmuramoyl-l-alanine amidase	0.21
CD0714	Conserved hypothetical protein	0.07
CD0758-CD0759 (plfAB)	Pyruvate formate-lyase	0.13-0.21
CD0888-CD0891 (speADEB)	Spermidine biosynthesis	0.08-0.12
CD0893	Iron-dependent hydrogenase	0.11
CD1695	Arsenical pump membrane protein	0.22
CD1768	Membrane protein	0.19
CD1893	Regulatory protein	0.19
CD2102	Na(+)/H(+) antiporter	0.05
CD2107	Permease	0.19	CD2107
CD2305	Pilin/general secretion pathway protein	0.23
CD2380 (iorB)	indolepyruvate oxidoreductase	0.22
CD2593	Antibiotic resistance ABC transporter, ATP-binding protein	0.23
CD3006	Probable alcohol dehydrogenase	0.17
CD3158	Conserved hypothetical protein	0.20	CD3158

^aA selection of genes, representing major functional classes, whose expression was affected ≥4-fold by a codY mutation. (All genes whose expression was affected ≥4-fold by a codY mutation are listed in Table S2 in the supplemental material.) Gene names reflect their positions in the strain 630 genome sequence, accompanied in some cases by the names of similar genes of known function in other organisms. Genes that lie in apparent operons are grouped.

^bPTS, phosphotransferase system; FMN, flavin mononucleotide.

^cFold change is the ratio of the hybridization signal for codY mutant RNA relative to codY⁺ RNA. The cutoff for significant change was set at 4-fold. Some genes, indicated by asterisks, had a fold change less than 4-fold but were included because they appeared to be in the same transcription unit with CodY-regulated genes.

^dGenes or transcription units that were identified by affinity purification of CodY-binding DNA fragments are indicated in the rightmost column.

The decision to include in our analysis only those genes for which the codY mutation caused a 4-fold or greater change in expression was arbitrary. We sought to restrict our attention to genes for which regulation by CodY has a significant biological impact. As a result, we may have underestimated the number of genes whose expression responds to CodY activity.

To confirm the gene expression results obtained from the microarray, both quantitative (Table 3) and semiquantitative (see Fig. S1 in the supplemental material) reverse transcription-PCR (RT-PCR) analyses were performed to determine transcript levels for selected genes in cells of strains JIR8094::pSD21 and JIR8094(pSD21) grown in TY medium to mid-exponential phase. The results of similar experiments were previously reported for the genes of the PaLoc (7). Consistent with the microarray results, transcript levels for ilvC, CD2344 (which encodes a putative membrane protein involved in γ-aminobutyrate metabolism), glgC, hisZ, CD1476 (encoding a putative cyclic di-GMP metabolism enzyme), and CD1803 (putatively encoding a cell surface protein) were found to be higher in the codY mutant (JIR8094::pSD21) than in the strains that expressed wild-type levels of CodY [JIR8094 and JIR8094(pSD21)] (Table 3 and Fig. S1). Also in accordance with the microarray results, speA transcript levels were found to be lower in the codY mutant than in the strains that expressed wild-type levels of CodY (Table 3 and Fig. S1). The apparent differential expression of the spoIIE gene seen in the microarray experiment, however, was not confirmed by qRT-PCR (Table 3).

TABLE 3.

Quantitative RT-PCR analysis of transcript levels

Gene	JIR8094 (pSD21) transcript data		JIR8094 codY::pSD21 transcript data
	Expression ratioa	P valueb	Expression ratioa	P valueb
ilvC	1.0 ± 0.3	0.86	85.9 ± 7.5	0.004
CD2344	1.3 ± 0.7	0.65	194.4 ± 103.2	0.11
glgC	1.7 ± 0.1	0.009	24.0 ± 5.4	0.027
hisZ	1.2 ± 0.2	0.33	17.7 ± 1.3	0.003
CD1276	1.5 ± 0.6	0.36	0.008 ± 0.0005	10−7
CD1476	0.9 ± 0.06	0.07	3.7 ± 0.9	0.048
CD1803	1.1 ± 0.07	0.39	53.6 ± 11.7	0.023
speA	1.3 ± 0.4	0.42	0.07 ± 0.04	0.0008
spoIIE	0.9 ± 0.17	0.40	1.5 ± 0.4	0.21

^aRNA from the codY⁺ and codY mutant strains was used as a template for cDNA synthesis. The cDNA was then amplified by PCR. The production of double-stranded DNA was monitored by detecting the binding of SYBR green by increased fluorescence. The amount of PCR product for each gene in each strain was normalized to the amount of an internal control gene, rpoC, in that strain and then compared to the ratio of rpoC to the gene in question in strain JIR8094 (set to 1.0 for all genes and not shown in the table). Each data point represents the average of triplicate assays of at least two RNA preparations.

^bThe P value was determined by a two-tailed Student's t test as a measure of the significance of the difference between expression in the tested strain and expression in strain JIR8094.

Genome-wide identification of direct CodY targets.

To identify genes that are directly activated or repressed by CodY, we performed a genomic DNA pulldown experiment using purified His₆-tagged CodY as bait, based on a method developed by Majerczyk et al. (27) (see Materials and Methods). A control experiment used as bait His₁₀-tagged B. subtilis aconitase, a protein that does not bind to DNA. The DNA fragments that bound to the His-tagged proteins were sequenced en masse. Using the sequencing results for unselected gDNA fragments as a baseline, we then determined a fold-enrichment value for each position on the chromosome by dividing the coverage map of each pulldown by the coverage map of the unselected gDNA. The coverage maps for the aconitase-containing sample and the unselected gDNA were not significantly different. Regions of DNA that were enriched more than 5-fold in the CodY-selected samples were compiled. Of the 628 regions identified, we selected the 350 most highly enriched CodY-binding regions for further characterization (see Table S3 in the supplemental material). Each of these regions had an enrichment factor of at least 17.

To assess the potential effects of CodY-DNA binding, we first surveyed the area of the chromosome surrounding each region of enrichment and determined the most likely regulatory targets of CodY. For example, the enriched region spanning coordinates 786094 to 786237 (based on NCBI sequence NC_009089.1) is located in the intergenic region upstream of the tcdR coding sequence. As no other transcripts are directed outwardly from this region, the most probable transcriptional target of CodY binding is tcdR, which encodes the sigma factor that regulates transcription of toxin genes (30). This region was previously shown to be a site of CodY binding by electrophoretic mobility shift and DNase I footprinting assays (7). Enriched regions near the tcdB and tcdC genes were also detected. The tcdB promoter region is a weak site of binding of CodY in gel shift assays (7), but the region upstream of tcdC includes a sequence to which CodY binds tightly in gel shift and DNase I footprinting experiments (data not shown). Other enriched regions correspond to many of the operons whose expression is affected by a codY mutation, including those for biosynthesis of the BCAAs (histidine, arginine, and glycogen), peptide, sugar and amino acid transport, ATP synthase, and butyrate metabolism. Interestingly, the spe operon, which is positively regulated by CodY, was not enriched in the pulldown experiment. Notably, 37 regions of enrichment were in or near regulatory genes, including those that encode transcription antiterminators, transcription activators, repressors, and two-component systems. Thus, it is likely that CodY influences the transcription of many transcripts indirectly by regulating the synthesis of other transcription factors.

Of the enriched regions, 118 contained one or more sequences with three or fewer mismatches to the consensus CodY-binding motif (Table 4 ). (The A+T-rich C. difficile genome contains only one site with the perfect consensus sequence, seven sites with one mismatch, 137 sites with two mismatches and 1,337 sites with three mismatches. Although the consensus sequence is clearly important for CodY binding [2], it is not usually sufficient [B. Belitsky, personal communication].)

TABLE 4.

Selected CodY-binding regions containing CodY consensus sequences

Genetic regiona	Coordinatesb	Potential CodY boxc	MMd	Potential CodY targete	Putative target function	Fold change in microarrayf
CD0013-CDs001	17426-17574	AAATTAAAGAAAATT/AATTCTAAAAAAATT	3/3	CDs001	Structural RNA	ND
CD0020-CD0021	34538-34706	AATTTTCACATAAAT	3	CD0021*	Pyruvate metabolism	0.74
CD0044-CD0045	69469-69716	AATTTTTAGAATTTT	3	CD0045	Carbohydrate metabolism	11.01
CD0107	136535-136773	ATTCTTCTAAAAATT	3	CD0108*	Nucleotide metabolism	0.87
CD111A-CD0112	155835-156107	AATATTTCGAATATT	3	CD0112*	Butanoate metabolism	0.65
CD0178	236362-236609	AATTTTATGAATATT	2	CD0179*	Glutamate metabolism	0.27
CD0214-CD0215	278547-278714	AATTATTTGAATATT	3	CD0214/CD0215*	Unknown/transposase	ND/0.49
CD0294-CD0295	356131-356235	AATATACTAAAAATT	3	CD0295*	Fe-S binding protein	1.25
CD0334-CD0335	407350-407557	AATATTCAGAAAATT	1	U
CD0351-CD0352	426229-426429	CATTTTCACAAAATT/AAATTTCAGAAAACT	2/2	CD0352*	Transcriptional regulator	7.47
CD0387-CD0388	457083-457305	TAGTATCAGAAAATT	3	CD0388*	Sugar transport	19.22
CD0440-CD0440A	523368-523530	AAGTTCCAGAAAAGT/AATTTTGTGAATATA	3/3	CD0440A	Transcriptional regulator	ND
CD0444-CD0445	528832-528971	AAAATTCAGAAAATT	2	CD0445*	Ornithine aminomutase	13.09
CD0536	640661-640900	AATATTCGGATTATT	3	CD0537*	Chemotaxis	0.81
CD0549-CD0550	654816-655019	AATATTCTGAAAAAT	2	CD0550	Unknown	0.88
CD0577-CD0578	690428-690792	AATTTTTATAAAGTT/AAATTAGAGAAAATT/TATTTTCTTAATATT	3/3/3	CD0577*	Phosphohydrolase	39
CD0658-CD0659	786357-786612	AATGTTGTCAAAATT/AATTTTCAAATAAAT	3/3	CD0659*	Transcriptional regulator	104.52
CD0664-CD0664A	805028-805140	TATTTTCCTAAAATA	3	CD0664	Toxin production	0.59
CD0669-CD0670	811707-811932	AATATTTTGAAAATA	3	CD0670	Transcriptional regulator	0.97
CD0678-CD0679	821392-821561	AATATTTAGAAAAAT	3	CD0679*	Unknown	0.76
CD0738-CD0739	901188-901315	TATTTTTAGAATATT/AATATTCAGTAAAAT	3/3	CD0738/CD0739	Unknown/unknown	0.80/1.16
CD0739-CD0740	902366-902610	TATTATAGGAAAATT	3	U
CD0781-CD0782	954316-954498	AATATTCTGTACATT	3	CD0781/CD0782	Penicillin binding protein/unknown	0.61/0.67
CD0799-CD0800	972046-972245	AATTTAGTGAAATTT	3	CD0800*	Crotonase	2.77
CD0812	985543-985711	AAAATTCTGAAAATT	2	CD0813	Sugar transport	0.89
CD0848-CD0849	1025023-1025198	AATTAACAAAAAATT	3	CD0849*	Unknown	7.69
CD0852-CD0853	1028194-1028302	AATTTGATGAAATTT	3	CD0852/CD0853*	Oligopeptide transport	0.95/12.59
CD0921	1100828-1101048	TATTCTTTGAAAATT	3	CD0921*	Phage protein	2.94
CD0988-CDs021	1152458-1152735	AATATTTAGAAAATT	2	CD0989*	Leucine biosynthesis	67.65
CD0989	1153010-1153269	ATTTTTAAGAAAATT	2	CD989*	Leucine metabolism	67.65
CD1031	1204553-1204718	GATATTCTGACAATT	3	CD1031*	Cellulose metabolism	0.53
CD1053-CD1054	1246326-1246760	AATTATTGAAAAATT	3	CD1054*	Butanoate metabolism	33.53
CD1071-CD1072	1265373-1265693	AATATTTAGAAAACT	3	CD1071/CD1072	Unknown/aminopeptidase	0.54/0.43
CD1084-CD1085	1277778-1277975	AATTTTATGAAAAAA/TATTTTTTGCAAATT	3/3	CD1085*	Glutamate metabolism	29.61
CD1120	1316544-1316724	CAGTTTCAGAAAATA	3	CD1120*	Glycerol metabolism	2
CD1211	1410549-1410682	AATATTCAGTAAATA	3	CD1211	Acetyltransferase	1.38
CD1259-CD1260	1463178-1463324	AATTTTTAGTTAATT/AATTTTCAGAAA	3/3	CD1260	Amino acid transport	23.7
CD1259-CD1260	1463631-1463859	AATATTGTGAAGATT	3	CD1260	Amino acid transport	23.7
CD1266	1471874-1472213	AATTTTCTTTTAATT	3	CD1266*	ABC transporter	9.27
CD1274-CD1275	1481454-1481685	AATTTTCATTAAATG	3	CD1275	*CodY*	1.15
CD1352	1569302-1569430	AATTCTGTGAAAACT	3	CD1352	ABC transporter system	0.49
CD1353	1569853-1570210	ATTTTTCCAAACATT	3	CD1353	Vitamin B6 metabolism	0.67
CD1378A-CD1378B	1595378-1595585	AATATTCAGAAGATG	3	CD1378, CD1378A-B	Transcriptional regulator/unknown	0.73/ND
CD1403	1624922-1625131	AATTTTCTGTATATA	3	CD1404*	Oligopeptide transport	0.36
CD1424-CD1425	1653848-1654028	ACTTGTCAGAAAATT	2	CD1424	Unknown	1.49
CD1511-CD1512	1749752-1749968	AATATTCAGTAAATA	3	CD1511*	Unknown	0.33
CD1546-CD1547	1795436-1795540	AATTTTTGAAAAATA	3	CD1547*	Histidine metabolism	15.06
CD1562-CDs028	1810325-1810537	AATAATCTGAAAATT	2	CD1563*	Structural RNA/BCAA metabolism	9.77
CDs031-CD1580	1831392-1831606	AAAAATCGGAAAATT	3	CDs031, CD1580*	Structural RNA/Amino acid metabolism	ND/1.99
CD1631	1890125-1890375	AATTTTGCAATAATT	3	CD1631*	Superoxide dismutase	0.8
CD1654	1918265-1918428	CATTTCCAGAAAATT	2	CD1654*	Lipoic acid metabolism	0.83
CD1663-CD1664	1936692-1936844	AAATTTATGAATATT	3	CD1663	Guanine deaminase	1.17
CD1703-CD1704	1978612-1978866	AATATTAAAAAAATT	3	CD1704*	Thiamine biosynthesis	2.29
CD1763-CD1764	2041381-2041551	AAATTTAAAAAAATT	3	CD1763/CD1764*	Unknown/unknown	2.21/1.78
CDs039-CD1774	2055675-2055882	AATATTTTAAAAATT/AATATTATGAAAAAT	3/3	CD1774*	Amino acid transport	3.69
CD1779	2060212-2060352	AATTTAAAGAAAATG/AATATACTCAAAATT	3/3	CD1779*	Unknown	1.27
CD1806-CD1807	2089882-2090196	AATTTTCTAATTATT/AATTTATAAAAAATT	3/3	CD1806	Fructokinase	0.77
CD1988-CDs044	2294168-2294473	AATTTTCAGTAAAAT	2	CD1988	Tryptophan transport	0.79
CD2007-CD2007A	2317619-2317709	AGTTTTGAGAATATT	3	CD2007	Erythromycin resistance	0.74
CD2010-CD2010A	2319981-2320126	AGTTTTGAGAATATT	3	CD2010*	Erythromycin resistance	0.51
CD2029-CD2030	2341391-2341584	AATTTTCATTAAAGT	3	CD2029*	Lipoprotein	0.57
CD2030-CD2031	2342425-2342606	TATTTTTTGAATATT	3	CD2030*	Ornithine carbamoyltransferase	10.58
CD2038-CD2039	2351975-2352131	AATTATCTGACAACT	3	CD2038*	Drug/Na+ antiporter	0.86
CD2108	2436840-2437027	TATTTTCTGATAATA	3	CD2108*	Antiporter	0.67
CD2109	2438222-2438382	TATTTACAGAAAATT	2	CD2109	Symporter	1.25
CD2123-CD2124	2459489-2459651	TTTTGTCAGAAAATT	3	CD2123/CD2124	Multi-protein complex assembly/transporter	0.67/0.89
CD2127-CD2128	2464269-2464411	AATTTTCACTATATT	3	CD2127	Unknown	10.72
CD2151-CD2152	2490473-2490640	AGTTTTCAGAACATT	2	CD2151/CD2152	Unknown/glutamyl aminopeptidase	2.64/38.05
CD2158-CD2159	2497849-2498061	AATTTTCAAAATATT	2	CD2158	Amino acid metabolism	31.98
CD2163-CD2164	2502982-2503060	AATTTTCTGAAATAA	3	CD2164	Metabolic pathways	6.78
CD2201-CD2202	2550642-2550769	TATATTCTCAAAATT	3	CD2201/CD2202	Sugar transport/symporter	0.47/1.30
CD2213-CD2214	2566366-2566491	AATTTTCAAAATATA	3	CD2213/CD2214	Nitrogen metabolism/transcriptional regulator	2.87/0.90
CD2217-CD2218	2569491-2569598	AATTTTCATAATATT	2	CD2217	Amino acid metabolism	1.43
CD2336-CD2337	2703228-2703567	TTTTTTCCGCAAATT	3	CD2336	Toxin anion resistance	0.84
CD2357-CD2358	2725674-2725854	TTTTTTCAGAAAATT	2	CD2357*	Pyrimidine metabolism	1.34
CD2418-CD2419	2791999-2792145	AATTTTCAGAAATAT	2	CD2418*	Sugar transport	0.48
CD2424	2799778-2800030	AATTTTCAGATAATA	2	CD2424*	Aminotransferase	27.37
CD2429A-CD2430	2805212-2805356	AATTTTAAGAATATT	2	CD2429*	Oxidoreductase	4.03
CD2442	2818120-2818383	AATTTTAGCAATATT	3	CD2441A	Phosphate starvation	ND
CD2450-CD2451	2825411-2825637	AATAGTCAGAATATT	3	CD2450*	Ribosomal methyltransferase	0.99
CD2459-CD2460	2838224-2838401	ATTTTTCAGAAAATT	1	CD2459*	Carbohydrate metabolism	0.7
CD2491-CD2492	2874398-2874563	ATATATCAGAAAATT	3	CD2492	Transcriptional regulator	0.93
CD2501-CD2502	2887393-2887548	TATTTTCATAATATT	3	CD2502	Amino acid metabolism	34.6
CD2510-CD2511	2899716-2899959	ACTTTTCTAAATATT	3	CD2511*	Transcription antiterminator/sugar transport	4.11
CD2517-CD2517A	2907057-2907203	AATTTTCAAACAATC	3	CD2517	Transcriptional regulator	0.99
CD2568-CD2569	2970536-2970719	AATTTTGGGAGGATT	3	CD2569	Carbohydrate metabolism	1.16
CD2617-CD2618	3024304-3024454	AAATTTTGGAATATT	3	CD2617*	Unknown	2.67
CDs055-CD2619	3027882-3028108	AATTCTCTGTAAAAT	3	CD2618	tRNA biosynthesis	1.36
CD2627-CD2628	3035283-3035389	ACTATTTAGAAAATT	3	CD2627*	Unknown	7.34
CD2675-CD2676	3091893-3092131	TTTATTCAGAAAATT	3	CD2675/CD2676	Transcriptional regulator/metabolism	0.70/22.89
CD2702-CD2703	3124233-3124541	TATTTTCAGAAAAAT	2	CD2702*	BCAA metabolism	6.78
CD2728-CD2729	3165395-3165598	AATCTTTTGAAAATC	3	U
CD2755-CD2756	3203460-3203634	ATTTTTTATAAAATT	3	CD2755*	Carbohydrate metabolism	1.34
CD2797	3261176-3261408	AATCTTCTGCTAATT	3	CD2796	Cell surface protein	0.72
CD2818-CD2819	3291095-3291200	ACTTTGCATAAAATT/AACTTATTGAAAATT	3/3	CD2819	Amino acid racemase	5.57
CD2827-CD2828	3300000-3300134	AATTTTTCGAAAATT	1	CD2827*/CD2828	Transcriptional regulator/aspartate aminotransferase	0.74/18.58
CD2836	3314798-3315060	AATTTTCTGATATTT	2	CD2835*	Aminobenzoyl-glutamate transport	29.28
CD2837-CD2828	3317022-3317154	AACTCTCAGTAAATT/AATTAAATGAAAATT	3/3	CD2837*	Unknown	3.99
CD2859	3343496-3343735	AATATTCTGTAAAAT/CATTTTCAGAAGAAT	3/3	CD2859*	d-Aminoacylase	4.28
CD2869	3356859-3357058	AATTATCAGAATTTT	3	CD2870	Sugar transport	2.7
CD2938-CD2939	3420861-3421088	AATTTTCAAAGAATA	3	CD2938*	Unknown	2.04
CD2947-CD2948	3425099-3425259	TATTTTCAAAATATT	3	CD2947*	Unknown	1.19
CD2957	3435338-3435581	AATATTCTCAAAATT	2	CD2957*	Electron transport	27.73
CD2960-CD2961	3438638-3438889	AATTTTCAGACATTC	3	CD2960*	Electron transport	28.32
CD2965	3443594-3443803	AGTTTTCTTATAATT	3	CD2965*	Signaling protein	6.79
CD2965	3444451-3444615	AGTTTTTGAAAAATT	3	CD2965*	Signaling protein	6.79
CD2983	3466329-3466483	AATTTTTTTATAATT	3	CD2983*	Unknown	2.33
CD3004-CD3005	3489854-3490056	AATTTTTTCAAACTT/AACTTTCTGATAATA	3/3	CD3004*/CD3005	Sugar transport/transcriptional regulator	0.19/1.38
CD3048	3543094-3543330	AAAATTCTGTAAATT/AATTTTTATAAAATC	3/3	CD3048*	Sugar transport	4.31
CD3138	3660297-3660539	AGATTTCATAAAATT	3	CD3138*	Transcriptional antiterminator	1.04
CD3157-CD3158	3688291-3688490	AATTTACTAAATATT	3	CD3157*/CD3158	SsrA-binding protein/unknown	1.08/0.20
CD3193	3736711-3736834	AATTTTCTTATAATA	3	CD3193	Phage-related protein	19.34
CD3223-CDs061	3772940-3773047	AATTTTAGCAATATT	3	CD3224*	Amino acid metabolism	0.91
CD3317-CD3319	3884259-3884475	AATTTTCTCACAATT	2	CD3317*	Metabolic pathways	3.12
CD3458-CD3459	4052622-4052839	ATTTTTCTGAATAGT	3	CD3458*	Unknown	12.05
CD3491	4082859-4083107	ACTTATCAGAAAATT	2	CD3491*	Exopolyphosphatase	0.71
CD3602	4209798-4209998	AAATTTCAAAAGATT	3	CD3602	ATPase	0.62
CD3628-CD3629	4236842-4237066	TATTATCTAAAAATT	3	CD3628*	Carbohydrate metabolism	6.17
CD3650	4258531-4258683	AATTTTTTGAAAGAT	3	CD3650*	Signaling protein	0.99

^aRegion of the strain 630 genome enriched by binding to CodY, as indicated by gene names. Of the 350 most-enriched regions, only those that contain CodY consensus sequences (≤3 mismatches) are included in this table. All of the 350 most-enriched regions are shown in Table S3 in the supplemental material.

^bCoordinates of the enriched region.

^cPutative CodY binding sequences in the enriched region, as defined by the consensus sequence (AATTTTCWGAAAATT) determined for B. subtilis and L. lactis. Up to three mismatches were allowed. Slashes separate multiple potential CodY boxes in the same region.

^dNumber of mismatches relative to consensus CodY box.

^eMost likely target of CodY binding based on location and orientations of neighboring genes. Genes with asterisks appear to be organized in multigene transcription units.

^fExpression ratio (codY mutant/wild type) for putative target genes.

The enriched sequences were also searched for a common motif as a means of testing the applicability of the CodY consensus binding sequence previously proposed for L. lactis and B. subtilis. As shown in Fig. 1, a motif found in 280 sequences had an overall E value of 5.2E−68 and included almost the entire previously derived consensus sequence, as well as an upstream sequence very reminiscent of the corresponding motif found for S. aureus (27).

FIG. 1.

Motif search in genomic regions enriched by binding to CodY. Using the MEME Suite, 293 regions of the C. difficile genome that had been enriched by affinity purification with His₆-tagged CodY were searched for a common motif. The resulting motif was found in 280 regions and proved to be similar to a motif discovered using the same methods for the S. aureus genome (27). Both motifs are compared to the CodY box defined for L. lactis (6, 10) and B. subtilis (2).

Of the 82 transcription units that were overexpressed in a codY mutant, 52 had enriched regions near the 5′ end of the transcription unit (those genes are indicated in Table 2; see also Table S2 in the supplemental material), suggesting that these genes are near sites of direct repression by CodY. Only two of the transcription units that were underexpressed in a codY mutant were associated with enriched regions (Table 2), suggesting that most CodY-activated genes are influenced by CodY indirectly.

Interaction of CodY with target promoter regions.

To confirm that CodY plays a direct role in the regulation of some of the genes identified as putative targets by both microarray and CodY pulldown screens, in vitro DNA binding experiments were performed. Radioactive double-stranded DNA probes corresponding to the upstream regions of the ilvC, CD2344, glgC, and hisZ genes were subjected to gel mobility shift analysis using purified C. difficile CodY (7). All four genes were highly overexpressed in the codY mutant strain (Table 2). C. difficile CodY bound to the DNA region upstream of all four genes, and this binding was enhanced in the presence of the known effectors of C. difficile CodY binding, GTP and a mixture of BCAAs (Fig. 2). The estimated K_D (equilibrium dissociation constant) for binding (approximated as the concentration of CodY that shifted 50% of the DNA [5]) in the presence of both effectors was 50 nM, 50 nM, <12.5 nM, and 12.5 nM CodY for the ilvC, CD2344, glgC, and hisZ DNA fragments, respectively. Based on previous analyses of the interaction of C. difficile CodY with DNA targets, the high binding affinities suggest that these DNA fragments are likely to contain direct targets of CodY in vivo (7). In contrast, only weak interactions (estimated K_D of >200 nM) were observed between C. difficile CodY and a DNA fragment containing the speA promoter region even in the presence of GTP and BCAAs (data not shown). These results confirm the pulldown results and indicate that the speADEB transcription unit is unlikely to be directly regulated by CodY.

FIG. 2.

Gel mobility shift assays for binding of CodY to selected gene promoter regions. DNA fragments containing the ilvC (423 bp), CD2344 (494 bp), glgC (780 bp), and hisZ (861 bp) promoter regions were incubated with increasing concentrations of CodY with and without the effectors GTP and branched-chain amino acids (BCAA), as described in Materials and Methods.

To identify specific binding sites for CodY, a DNase I footprinting assay was used to analyze the interaction of CodY with the DNA fragments for which CodY had high affinity. Since the specific interaction of CodY with ilv gene promoters has been extensively studied in other Gram-positive bacteria (2, 36, 45, 50), we focused on the CD2344, glgC, and hisZ regulatory regions. The binding site in the hisZ region with the highest affinity was located at positions −588 to −625 relative to the ATG start codon (Fig. 3). A 15-bp sequence, AATTcTCTGAAAcac, with four mismatches (indicated by lowercase letters) compared to the consensus CodY-binding motif, overlaps with the protected region. The transcription start point of the hisZ gene, as determined by 5′ RACE analysis, however, lies about 300 bp downstream of the high-affinity CodY binding site. How CodY regulates hisZ from the observed site is unclear.

FIG. 3.

DNase I protection assays for binding of CodY to selected gene promoter regions. Binding reaction mixtures contained 2 mM GTP, 10 mM BCAA, and various concentrations of CodY (nM) as described in Materials and Methods. Protected regions are marked by vertical bars. The location of the transcription start site for each gene is indicated by a bent arrow that also indicates the direction of transcription.

For the fragment of DNA containing the CD2344 promoter region, a region that spans the sequence from positions −310 to −350 relative to the ATG start codon was protected by CodY from DNase I digestion in the presence of GTP and BCAAs, with protection appearing at 100 nM CodY (Fig. 3). This region of protection was fully contained within the region between CD2344 and CD2345 that was highly enriched in the pulldown experiment (Fig. 4). To correlate the location of the CodY binding site with the transcription start site, a primer extension experiment was performed. Primer extension analysis revealed an apparent transcription start located 300 bp upstream of the CD2344 initiation codon (Fig. 5). The predicted −10 and −35 promoter sequences would lie within the CodY binding region as defined by DNase I protection, fitting a model in which CodY acts as a direct repressor of CD2344 transcription. It is worth noting that the primer extension product was detected only when RNA from the codY mutant strain was used, affirming that the transcript in question is regulated by CodY (Fig. 5). A 15-bp sequence reminiscent of the CodY consensus site, AAgTTTtTGAtAtTT, with four mismatches (indicated by lowercase letters), was identified within the CodY protected region and overlapping with the putative promoter site. Interestingly, CD2345, the gene immediately upstream of CD2344 (Fig. 4), is transcribed divergently with respect to CD2344 and encodes a protein similar to LysR family regulatory proteins. LysR regulator genes are frequently transcribed divergently from the first genes of the operons they control. We found no evidence that CD2345 is regulated by CodY.

FIG. 4.

Regulatory site architecture. The locations of transcription start sites (+1), promoter −10 and −35 elements, start codons (ATG), CodY binding sites determined by DNase I protection assays (PR; gray blocks), and regions enriched in CodY affinity chromatography experiments (hatched blocks) are indicated for three C. difficile regulatory regions. Coordinates are given with respect to the transcription start site. The regions of the tcdR locus protected by CodY were determined in previous work (7).

FIG. 5.

Primer extension analysis to map the start sites for the CD2344 and glgC transcripts. Oligonucleotide primers radioactively labeled with ³²P at the 5′ end were annealed to total RNA extracted from the codY mutant, JIR8094::pSD21, and the codY⁺ strain, JIR8094 (pSD21), as described in Materials and Methods. Arrows indicate the positions of the major CodY-regulated primer extension products.

In a DNase I footprinting experiment for the glgC promoter region, CodY protected a region from positions −595 to −629 relative to the ATG start codon, with protection beginning to appear at 25 nM CodY (Fig. 3). The region of protection overlapped with the region that was enriched in the pulldown experiment. A primer extension experiment established that the glgC transcription start point appears to be 560 bp upstream of the initiation codon (Fig. 5). The −35 promoter sequence for this transcript would be predicted to overlap with or be directly downstream of the CodY binding region (Fig. 4), fitting a model in which CodY would act as a repressor of glgC gene expression. As seen for the CD2344 primer extension experiment, the glgC primer extension product was detected only when codY mutant strain RNA was used as a template, indicating that the transcription start site identified is for a transcript regulated by CodY (Fig. 5). Two 15-bp sequences, AATTaaCAGAAtcTT and AATcTTtAtAcAATT, each with four mismatches compared to the consensus CodY-binding motif, were identified within the CodY protected region.

The role of CodY in virulence gene regulation.

In previous work, we showed that CodY is a repressor of tcdR, the gene that encodes the sigma factor that is needed for expression of the major virulence factors of C. difficile. We found three binding sites for CodY in the region upstream of the tcdR gene (7) but were unable to confidently predict the effect of CodY binding on gene expression because we had yet to reliably map the transcription start site for the tcdR gene. By integrating new information from the pulldown experiment and 5′ RACE, we were able to orient the binding sites with respect to the transcription start point. For 5′ RACE analysis, RNA was isolated from mid-exponential-phase cells of the codY mutant strain JIR8094::pSD21, a strain that expresses high levels of tcdR transcript. The 5′ end of the tcdR transcript mapped to an A residue located 16 bp upstream of the ATG start codon (data not shown). The −35 promoter sequence for this transcript would be predicted to fall within the previously determined CodY-protected region III (Fig. 4) (7). Two 15-bp sequences, AATgTTgTcAAAATT and AATTTTCAaAtAAaT, each with three mismatches compared to the consensus CodY-binding motif, lie within or overlap with the CodY-protected region. Two other CodY-protected regions (I and II) lie further upstream (7). These latter protected regions fall within or overlap with the region enriched in the pulldown experiment (Fig. 4).

The tcdR gene is very likely to be transcribed from two promoters, one recognized by the σ^A-containing form of RNA polymerase and a second promoter that is recognized by the TcdR-containing RNA polymerase. The σ^A-dependent promoter should be active at least at a low level during the exponential growth phase to allow basal expression of tcdR (TcdR cannot be totally dependent on itself for its synthesis), whereas the TcdR-dependent promoter should be activated only in stationary phase when toxin synthesis is maximal. It would be reasonable to suggest that the CodY-protected regions I and II correspond to one of the promoters and the protected region III corresponds to the other. But which is which? The CodY binding sites I and II and the region of enrichment in the pulldown experiment lie within or downstream of a TcdR-type promoter sequence (31), whereas site III overlaps with a σ^A-type promoter sequence. Thus, the most likely scenario is that CodY binds to both promoters, preventing transcription from both during rapid exponential growth. Note that the 5′ RACE results revealed only one transcription start, corresponding to the σ^A-type promoter. When two transcripts overlap, however, 5′ RACE preferentially amplifies the shorter transcript.

Whereas the synthesis of the major toxins A and B is regulated by CodY, the synthesis of the third toxin, CDT, is not. We found no evidence for altered expression of the cdtAB operon in a codY mutant and no enrichment for the cdt locus in the pulldown experiment. Other potential virulence factors do appear to be under CodY control, however. CD0440 and CD1803 encode putative surface proteins that might interact with host cells. The genes CD0757 and CD1476 were highly expressed in the codY mutant; the CD1476 region was also identified as a site of CodY binding, as were the genes CD2385, CD2873, CD2965, and CD3650. All of these genes are predicted to have EAL and/or GGDEF domains, which are associated with the synthesis and breakdown of the second messenger molecule, cyclic di-GMP (39). In other bacteria, cyclic di-GMP has been implicated in control of motility, biofilm formation, and virulence gene expression (49). If CodY affects signaling through cyclic di-GMP, its influence on virulence could be considerably broader than we currently know.

The coregulation of metabolic genes and virulence genes by CodY underscores the connection between nutritional sufficiency and pathogenesis in C. difficile. This bacterium has evolved to express its most potent virulence factors at high level only after experiencing nutrient deprivation. Given the wide occurrence of CodY homologs in Gram-positive pathogens and their implication in virulence in several such species, this strategy appears to be widespread.