Identification of CodY Targets in Bacillus anthracis by Genome-Wide In Vitro Binding Analysis

Abstract

In Gram-positive bacteria, CodY is an important regulator of genes whose expression changes under conditions of nutrient limitation. Bacillus anthracis CodY represses or activates directly or indirectly approximately 500 genes. Affinity purification of CodY-DNA complexes was used to identify the direct targets of CodY. Of the 389 DNA binding sites that were copurified with CodY, 132 sites were in or near the regulatory regions governing the expression of 197 CodY-controlled genes, indicating that CodY controls many other genes indirectly. CodY-binding specificity was verified using electrophoretic mobility shift and DNase I footprinting assays for three CodY targets. Analysis of the bound sequences led to the identification of a B. anthracis CodY-binding consensus motif that was found in 366 of the 389 affinity-purified DNA regions. Regulation of the expression of the two genes directly controlled by CodY, sap and eag, encoding the two surface layer (S-layer) proteins, was analyzed further by monitoring the expression of transcriptional lacZ reporter fusions in parental and codY mutant strains. CodY proved to be a direct repressor of both sap and eag expression. Since the expression of the S-layer genes is under the control of both CodY and PagR (a regulator that responds to bicarbonate), their expression levels respond to both metabolic and environmental cues.

INTRODUCTION

The Gram-positive spore-forming pathogen Bacillus anthracis is the causative agent of anthrax. Anthrax is primarily a disease of herbivores, but it can also be found in humans and other mammals (1). B. anthracis spores initiate infection. Once inside the host, the spores germinate, generating vegetative cells that multiply in host tissues. It is during this proliferative stage that B. anthracis produces its key virulence factors, namely, a tripartite toxin and a poly-γ-d-glutamate capsule, which are encoded on the virulence plasmids pXO1 and pXO2, respectively. The production of toxin and capsule responds to environmental cues, such as the presence of CO₂ and bicarbonate and a temperature of 37°C (2, 3). The production of toxin components is low during exponential growth and reaches its highest levels during entry into the stationary phase (3, 4). The regulator protein AtxA, which is encoded by pXO1, is essential for the expression of toxin components, and, during in vitro growth, it influences the expression of the capsule (cap) operon (4–7).

The global regulatory protein CodY influences the expression of many metabolic and virulence genes in Gram-positive pathogens in response to the availability of GTP and the branched-chain amino acids (isoleucine, leucine, and valine) (8, 9). In B. anthracis, CodY indirectly activates toxin gene expression, apparently by influencing AtxA stability (10, 11). In contrast, the deletion of codY does not affect B. anthracis capsulation (12). Evidently, the dramatically lowered AtxA levels in the codY mutant do not result in decreased expression of the capsule operon.

Genome-wide transcriptional profiling of the codY mutant strain indicated that a codY-null mutation affected the expression of >200 genes during both the exponential and early stationary phases and approximately 500 genes during at least one of these two growth phases (11). CodY acts as a negative regulator of many of these genes and activates the expression of other genes, including those encoding iron uptake and detoxification systems, in addition to the toxin genes (11). CodY is needed for the uptake of organic iron compounds, and consequently, the deletion of codY may lead to an in vivo growth defect when iron becomes the limiting factor for growth (12).

The transcription profiling experiment revealed that sap and eag, which encode the components Sap and EA1, respectively, of the surface layer (or S-layer) (13, 14), are among the genes repressed by CodY. The S-layer of B. anthracis is a proteinaceous paracrystalline sheath that forms the outermost surface structure of noncapsulated vegetative cells (15). The function of the S-layer of B. anthracis is not completely clear; it might act as a molecular sieve or it might protect the bacterium from the actions of complement (16). However, the deletion of B. anthracis S-layer-encoding genes has no discernible effect on virulence (17). In cells grown in rich medium, sap and eag are temporally regulated. In addition, Sap represses eag transcription (18). In contrast, during the growth of B. anthracis inside the host or in a defined medium containing bicarbonate, eag is expressed at higher levels than sap (14, 19). This differential regulation is due to the indirect action of AtxA on the expression of these genes. The regulator PagR, which binds to the promoter regions of both sap and eag, acts as a repressor of sap and an activator of eag transcription. The pagR gene is the second gene of the AtxA-dependent pagAR operon (19–21).

The in silico analysis of our global transcription results showed that many CodY-controlled genes are associated with a sequence reminiscent of the consensus CodY-binding site first identified in Lactococcus lactis (22, 23); this implies that these genes are direct targets of CodY-mediated regulation. To determine more rigorously which genes are the direct targets of CodY, we used an affinity purification assay (24, 25) and extended this analysis by using gel shift, DNase I footprinting, and lacZ fusion experiments for three target genes. Our results reveal >130 CodY-binding sites that control the expression of 197 genes, as well as additional CodY-binding sites that either control genes under conditions not yet tested or play no role in transcriptional control. The sap and eag genes proved to be among the direct targets of CodY. The sites of CodY binding in their regulatory regions identified by affinity purification were confirmed by DNase I footprinting. Moreover, the transcriptional regulation of sap and eag by CodY was confirmed by an analysis of lacZ transcriptional fusions.

MATERIALS AND METHODS

Strains and growth conditions.

The B. anthracis strain RTC10 (pXO1 lef686 cya346-353, pXO2 ΔcapB) (26) was used to prepare the DNA for the pulldown experiments. The strains harboring lacZ transcriptional fusions were used to assay transcription initiating from the relevant promoters: 7SZ (sap-lacZ) and 7EZ (eag-lacZ) and their codY derivatives, 7SZC and 7EZC (19 and this study). All work with B. anthracis strains was carried out in accordance with the biosafety guidelines mandated by the country in which each assay was performed.

Growth of the cultures was followed by measuring the absorbance at 600 nm (A₆₀₀). B. anthracis precultures were grown overnight at 37°C in BHI (brain heart infusion) (Difco) medium supplemented with 0.5% glycerol, with rotary shaking at 150 rpm. Experimental cultures were grown at 37°C under a 5% CO₂ atmosphere in R medium (27) supplemented with 0.6% sodium bicarbonate (RBic) (2). RBic was inoculated from an overnight culture to a starting A₆₀₀ of 0.05. Aeration of the culture was achieved by continuous stirring with a magnetic bar at 450 rpm. For DNA extraction, the culture was grown at 37°C in BHI until it reached mid-exponential phase (A₆₀₀ = 0.8). The antibiotics used were spectinomycin (100 μg/ml) and kanamycin (40 μg/ml).

Generation of codY mutants in B. anthracis.

A null mutation in the codY gene generated by the insertion of a spectinomycin resistance cassette at codon 77 of codY has been described (11). The mutation was transferred by filter mating using E. coli HB101(pRK24)(pCOD30) as the donor and B. anthracis strain 7SZ or 7EZ as the recipient (19, 28, 29).

Overexpression and purification of B. anthracis CodY.

E. coli strain JM107(pJP15) (11) was grown in Luria broth containing ampicillin (50 μg/ml) to an A₆₀₀ of 0.7. Expression of the codY gene was induced by the addition of arabinose to a final concentration of 0.2%. After 4 h of further incubation, the cells were harvested and CodY-His₆ was purified using a previously described method (30). Elution from the Talon Co²⁺ metal affinity column (Clontech) was carried out using a buffer containing 20 mM Tris-HCl (pH 8.0), 125 mM KCl, 5 mM β-mercaptoethanol, 10% (vol/vol) glycerol, 1 mM phenylmethylsulfonyl fluoride, 2 μM pepstatin-A, and 75 mM imidazole. The purified B. anthracis CodY protein was free of contaminating proteins as tested by SDS-PAGE and was stored in 20-μl aliquots at −80°C. For each experiment, one tube of protein solution was removed from the freezer, thawed, and diluted as needed.

Affinity purification of CodY-DNA complexes and analysis using the Illumina genome analyzer II.

Genomic DNA was extracted from B. anthracis RTC10 as described in reference 31. Four samples of 10 μg of DNA each were dissolved in 500 μl of sterile deionized water and sheared by sonication. The method used was initially developed by Majerczyk et al. (25) and improved by Dineen et al. (24). In brief, adapters were ligated to the ends of the sheared DNA fragments and DNA fragments of 300 to 500 bp were purified. Amplification of the fragments was performed by PCR using primers that were complementary to the adapter sequences. The DNA fragments were incubated with purified His₆-tagged B. anthracis CodY (50 or 100 nM) in the presence of the coeffectors, 5 mM GTP and 10 mM each isoleucine, leucine, and valine. A negative control consisted of carrying out the same experiment using purified Bacillus subtilis His₁₀-aconitase (32) in place of B. anthracis CodY. Purification of CodY that was bound to its target genes was performed on a Talon Co²⁺ metal affinity resin (Clontech). The DNA fragments that copurified with the His-tagged proteins were sequenced bidirectionally by the Tufts University Nucleic Acid and Protein Core Facility using an Illumina genome analyzer II.

The sequence data were analyzed in two ways. First, the sequences of the first 40 bp and last 40 bp of each DNA fragment were determined directly. Then, for each fragment, the sequence between the ends was filled empirically based on the reference sequence of the B. anthracis Sterne strain chromosome (NCBI accession no. NC_005945) and plasmids pXO1 (NCBI accession no. NC_007322) and pXO2 (NCBI accession no. NC_007323). Using the bioinformatics tools available through the Galaxy Project (33–35), the sequence data were filtered for quality using the NGS QC algorithm with a quality cutoff of 7 for 90% of the bases. About 10% of reads were discarded as being of low quality. The remaining reads were aligned against the B. anthracis Sterne genome sequence and pXO1 and pXO2 sequences using the next-generation sequencing (NGS) mapping tool Bowtie (settings, trim 5 nucleotides [nt] from the 3′ end; allow 3 mismatches; quality value = 70; seed length = 23; “best” option). The resulting alignments were then analyzed using Hop-Count, which plots the number of times each nucleotide appears as the first base after the adapter at the 5′ end of the DNA fragment being sequenced. Hop-Count is a script written by K. Bodi and D. Lazinski for the Tufts Nucleic Acid and Protein Core Facility. Following an approach devised by B. Belitsky (unpublished data), the Hop-Count output was manually scanned for regions in which reads appeared on the top strand but not on the bottom strand, followed by a region with counts on the bottom strand but not on the top strand. In >90% of cases, the bases at the junction between the regions on the top and bottom strands had few, if any, reads and were taken to define the nucleotides essential for CodY binding. For the data compilation reported in Table S1 in the supplemental material, only regions detected in at least two independent experiments were included.

In silico analysis.

Of the 389 identified target sequences, 375 binding sites that consisted of ≥8 bp were searched for conserved motifs using the MEME suite v4.8.1 (36). The parameters were as follows: 0 to 1 occurrence per fragment on either strand of DNA and a width of 6 to 50 nt.

Gel mobility shift and DNase I footprinting assays.

For gel mobility shift assays, DNA probes were amplified by PCR using a 5′-end ³²P-labeled primer, or the PCR product was labeled after amplification. Methods for primer labeling, binding of CodY to DNA, and electrophoresis of DNA-protein complexes in nondenaturing polyacrylamide gels have been described (30). To locate CodY-binding sites precisely, PCR products labeled at the 5′ end of only one strand were incubated with CodY, treated with RQ1 RNase-free DNase (Promega), and subjected to electrophoresis in urea-polyacrylamide gels as described previously (30, 37). A DNA sequence ladder was run alongside the DNase I digestion products. The ladder was prepared by the Sanger method, using the Sequenase kit (United States Biochemical) and [α-³⁵S]dATP, as recommended by the manufacturer. The primers used were oPJ37 (GGTATGGTAATTTTGAAAAATCTAACGCTA) and oPJ38 (CGACAAAAAAGTGAATTTTCACAATTCTAA) for the sap gene, oPJ36 (CCTCCTTCAGGAATATGCTACTAGTTTTAC) and oPJ33 (GCGCTTTCAAAAATAGAAGAATACCTCATT) for the eag gene, and oPJ29 (GCTATTTCAATAGAAGAAACAAAAAACCAA) and oPJ44 (CGATGGATATCGGTGTTAGCATGTC) for the atxA gene.

β-Galactosidase assays.

The accumulation of β-galactosidase by the various lacZ fusion strains was determined as described previously (3) in samples taken during growth in RBic at 37°C in a 5% CO₂ atmosphere. All experiments were performed at least in triplicate; the representative results are shown below (see Fig. 4).

Fig 4.

CodY affects transcription of sap and eag. B. anthracis 7EZ (eag-lacZ) and 7SZ (sap-lacZ) (circles) and their respective codY mutants (squares) were grown in RBic medium under a 5% CO₂ atmosphere. Growth was followed by A₆₀₀ measurements (open symbols), and the transcription of sap and eag was determined by measuring β-galactosidase activity (closed symbols). The results shown are representative of three independent experiments.

RESULTS

Genome-wide identification of direct CodY targets.

Previous genome-wide transcriptional profiling showed that >200 genes are differentially expressed during the exponential and early stationary growth phases and that >500 genes are differentially expressed during at least one of these growth phases in a codY-null mutant of B. anthracis compared to in its parent strain (11). However, this analysis did not distinguish between direct (binding of CodY to target promoters) and indirect regulation of transcription. To identify genes that are directly controlled by CodY, we performed an affinity purification experiment using purified His₆-tagged B. anthracis CodY as bait, based on a method developed by Majerczyk et al. and Dineen et al. (24, 25). Three hundred eighty-nine CodY-binding regions were detected in at least two of three independent experiments (see Table S1 in the supplemental material). In 16 cases, more than one region was identified upstream of a single gene. Of the 389 regions identified (see Table S1), 132 regions were within the putative regulatory regions of 131 transcription units, containing 197 genes (Table 1) whose expression was shown previously to be affected by the disruption of codY (11). Thus, all of these genes appeared to be under the direct control of CodY. Of these 197 genes, 182 were repressed by and 15 were activated by CodY. The repressed genes include those encoding proteins involved in amino acid biosynthesis and transport, as well as nucleic acid metabolism, the pentose pathway, the tricarboxylic acid cycle, acetyl coenzyme A synthesis, and oxidative stress functions. Nine genes coding for regulatory proteins appeared to be repressed directly by CodY, including nprR and sinR. Furthermore, CodY bound to sites upstream of and within BAS0093, the gene that encodes the RNA polymerase sigma factor σ^H. Our analysis also revealed that the S-layer genes sap and eag are direct targets of CodY. Transcriptional profiling previously indicated that CodY represses the expression of these genes (11).

Table 1.

Regions of the B. anthracis genome selected by in vitro binding of CodY

Coordinates (start to end)a	Position relative to coding sequence	Potential targetb,c gene	Putative target gene function	Fold-change microarrayd,e
				A600 = 0.2	A600 = 1.0
Chromosome
15371–15390	Upstream	BAS0011 (guaB)	IMP dehydrogenase	6.5	4.7
41312–41347	Upstream	BAS0036 (abrB)	Transition state regulator	ND	0.41
73440–73533	Upstream	BAS0067 (cysK-1)	Cysteine synthase A	ND	0.36
102880–102952	Upstream	BAS0093 (sigH)	RNA polymerase sigma-H factor	0.29	0.6
103459–103483	Within	BAS0093 (sigH)	RNA polymerase sigma-H factor	0.29	0.6
218863–219056	Upstream	BAS0218	Oligopeptide ABC transporter substrate-binding protein	ND	0.13
226686–226720	Upstream	BAS0226 (hpp)*	4-Hydroxyphenylpyruvate dioxygenase	0.07	0.13
230217–230315	Upstream	BAS0229	Major facilitator family transporter	0.37	0.37
338219–338258	Upstream	BAS0314	Aminopeptidase AmpS	0.33	ND
525660–525695	Upstream	BAS0498	Glutamate synthase, large subunit	0.25	0.46
548451–548475	Upstream	BAS0518	Hypothetical protein	0.2	0.43
611816–612018	Upstream	BAS0566 (nprR)*	Transcriptional regulator	0.19	0.33
613543–613558	Upstream	BAS0567 (nprB)	Neutral protease	0.15	0.29
628955–628973	Upstream	BAS0580	Proton/peptide symporter family protein	17.9	13.6
660284–660452	Upstream	BAS0610	Amino acid ABC transporter permease	0.46	0.47
688599–688933	Upstream	BAS0638 (inhA2)	Immune inhibitor A metalloprotease	0.06	0.09
714089–714124	Upstream	BAS0660	Xanthine/uracil permease family protein	4.3	4.3
723269–723285	Upstream	BAS0669	Quinol oxidase subunit II	ND	2.5
831170–831178	Upstream	BAS0779	Amino acid permease	0.04	ND
859284–859309	Upstream	BAS0809	ABC transporter ATP-binding protein/permease	0.19	ND
896336–896512	Upstream	BAS0841 (sap)	S-layer protein, Sap	0.18	ND
899355–899413	Upstream	BAS0842 (eag)	S-layer protein, EA1	0.04	ND
899525–899553	Upstream	BAS0842 (eag)	S-layer protein, EA1	0.04	ND
909509–909607	Upstream	BAS0847	Hypothetical protein	0.26	ND
	Upstream	BAS0848*	Enoyl-CoA hydratase	0.34	ND
1075667–1075805	Within	BAS1022	Wall-associated protein	11.9	ND
1141119–1141158	Upstream	BAS1089	Hypothetical protein	0.33	ND
1160748–1160829	Upstream	BAS1107	Oligopeptide ABC transporter substrate-binding protein	7.5	11
1240253–1240257	Upstream	BAS1196 (sinI)*	SinI protein	0.14	0.28
	Upstream	BAS1197 (inhA1)	Immune inhibitor A metalloprotease	0.01	0.01
1262756–1262940	Upstream	BAS1216	Hypothetical protein	0.39	ND
1272738–1272770	Upstream	BAS1228*	PadR family transcriptional regulator	0.41	ND
	Upstream	BAS1229*	PhaR protein	0.07	0.18
1296547–1296595	Upstream	BAS1255	Hypothetical protein	0.07	0.40
1316059–1316118	Upstream	BAS1281	Hypothetical protein	0.31	ND
1327364–1327404	Upstream	BAS1293	Bla/Mec family transcriptional regulator	0.22	0.35
1339040–1339093	Upstream	BAS1307*	Branched-chain amino acid aminotransferase	0.01	0.01
1345169–1345198	Upstream	BAS1312	3-Isopropylmalate dehydrogenase	0.01	0.01
1350300–1350370	Within	BAS1316 (hisG)	ATP phosphoribosyltransferase	ND	0.28
1408429–1408467	Upstream	BAS1376*	Hypothetical protein	0.27	0.31
1408807–1408869	Upstream	BAS1377	Hypothetical protein	0.38	0.46
1414452–1414476	Upstream	BAS1384	ResB protein	0.34	0.42
1544764–1544827	Upstream	BAS1512*	Cold shock protein CspB	ND	3.8
1547523–1547563	Upstream	BAS1517	Hypothetical protein	ND	0.34
1594231–1594417	Upstream	BAS1573	Hypothetical protein	0.44	0.41
1730383–1730439	Upstream	BAS1713	Branched-chain amino acid aminotransferase	0.13	0.08
1731059–1731100	Upstream	BAS1713	Branched-chain amino acid aminotransferase	0.13	0.08
1737014–1737109	Within	BAS1718	Threonine dehydratase	0.10	0.06
1784227–1784259	Upstream	BAS1763	Xaa-Pro aminopeptidase	0.28	0.34
1845838–1845875	Upstream	BAS1820	Aminoglycoside 6-adenylyltransferase	0.45	ND
1850239–1850264	Upstream	BAS1825*	Homoserine dehydrogenase	ND	0.28
1851801–1851869	Within/upstream	BAS1825*/BAS1826	Homoserine dehydrogenase/threonine synthase	ND	0.28/0.31
1878275–1878294	Upstream	BAS1853*	Hypothetical protein	ND	0.34
1887952–1888031	Upstream	BAS1866	Ribosome-associated GTPase	0.33	0.45
1894901–1894936	Upstream	BAS1873	Hypothetical protein	ND	0.32
1916527–1916633	Upstream/start	BAS1895	Fosfomycin resistance protein FosB	ND	0.49
1933891–1933922	Upstream	BAS1916	Branched-chain amino acid transport system II carrier protein	ND	0.35
1946712–1946796	Upstream	BAS1932	Alanine racemase	ND	0.38
1980054–1980077	Upstream	BAS1975*	Hypothetical protein	0.06	0.06
2076907–2075914	Upstream	BAS2067	Aspartate racemase	0.40	0.39
2103632–2103712	Upstream	BAS2100	d-Amino acid aminotransferase	0.19	0.25
2115360–2115409	Upstream	BAS2113	Hypothetical protein	0.36	ND
2122151–2122209	Upstream	BAS2122*	Azoreductase	0.18	0.32
2182362–2182473	Upstream	BAS2188	Citrate synthase 3	0.01	0.02
2211175–2211213	Within	BAS2208	Nonribosomal peptide synthetase DhbF	33.3	ND
2260117–2260143	Upstream	BAS2261	HAD superfamily hydrolase	ND	0.29
2362344–2362368	Upstream	BAS2360*	Hypothetical protein	0.17	ND
2370310–2370337	Upstream	BAS2369*	Acyl-CoA dehydrogenase	0.11	0.36
2567845–2568001	Upstream	BAS2567*	Kynureninase	0.31	0.34
2569082–2569180	Upstream	BAS2569	MutT/Nudix family protein	0.49	0.45
2649951–2649973	Upstream	BAS2664	Hypothetical protein	0.24	ND
2652172–2652185	Upstream	BAS2667	X-prolyl-dipeptidyl aminopeptidase	0.04	0.05
2701105–2701226	Upstream	BAS2722*	Aminoglycoside N3-acetyltransferase	0.38	0.48
2765653–2765696	Upstream	BAS2785	IclR family transcriptional regulator	0.32	0.30
2765754–2765801	Upstream	BAS2785	IclR family transcriptional regulator	0.32	0.30
2809027–2809179	Upstream	BAS2830	MutT/Nudix family protein	0.05	0.41
2811584–2811656	Upstream	BAS2835*	GntR family transcriptional regulator	0.03	0.37
2956380–2956449	Upstream	BAS2983	Glyoxylase	0.35	0.49
3010767–3010790	Upstream	BAS3037*	Hypothetical protein	0.01	0.01
	Upstream	BAS3038	LysR family transcriptional regulator	ND	0.09
3053679–3053700	Upstream	BAS3072	Isochorismatase family protein	0.24	0.34
3060852–3061055	Upstream/start	BAS3079*	Phosphoserine aminotransferase	0.15	0.16
3201993–3202017	Within	BAS3228	Glycosyl transferase family protein	2.7	ND
3280046–3280069	Upstream	BAS3315*	Glycolate oxidase subunit GlcD	0.07	0.05
3318842–3318863	Upstream	BAS3348	Aldehyde dehydrogenase	0.09	0.26
3354721–3354753	Upstream	BAS3379	Oligopeptide ABC transporter substrate-binding protein	0.18	0.19
3381975–3382037	Within	BAS3408	Aconitate hydratase	0.30	ND
3394054–3394086	Upstream	BAS3421	Medium-chain-fatty-acid-CoA ligase	0.05	0.06
3395833–3395844	Upstream	BAS3424	Hypothetical protein	ND	3.6
3417205–3417232	Upstream	BAS3443 (hutP)	Antiterminator HutP	0.15	0.40
3647195–3647222	Within	BAS3682 (xerC)	Site-specific tyrosine recombinase XerC	ND	0.54
3647543–3647670	Upstream	BAS3682 (xerC)	Site-specific tyrosine recombinase XerC	ND	0.54
3800292–3800414	Upstream	BAS3850	Hypothetical protein	ND	0.43
3801987–3802007	Upstream	BAS3851 (amiF)*	Formamidase	0.03	0.05
3805046–3805095	Within	BAS3855	Cytochrome c oxidase subunit I	0.16	0.44
3806497–3806525	Within	BAS3856	Cytochrome c oxidase subunit II	0.17	0.41
3822014–3822030	Upstream	BAS3871*	Hypothetical protein	0.10	0.13
3826974–3827044	Upstream	BAS3878	Hypothetical protein	ND	0.47
3843179–3843223	Upstream	BAS3894	Hypothetical protein	0.32	ND
3847974–3848000	Upstream	BAS3900	Short chain dehydrogenase	0.13	0.29
3859657–3859902	Upstream	BAS3912	5- Methyltetrahydropteroyltriglutamate-homocysteine S-methyltransferase	0.13	0.07
3865593–3865622	Upstream	BAS3916	Sporulation kinase B	0.12	0.16
	Upstream	BAS3917	Hypothetical protein	0.37	ND
3878080–3878289	Upstream	BAS3926	Hypothetical protein	ND	0.42
3924839–3924870	Upstream	BAS3982*	Sodium-dependent symporter family protein	0.02	0.02
3946199–3946351	Upstream	BAS4009	5′-Nucleotidase	0.14	0.17
3965600–3965617	Upstream	BAS4029	Hypothetical protein	0.15	0.31
4001362–4001531	Within	BAS4069	Butyrate kinase	0.37	ND
4058043–4058167	Within	BAS4138	Prolyl 4-hydroxylase subunit alpha	7.4	6.8
4063390–4063526	Upstream	BAS4147	Hypothetical protein	0.19	ND
4078047–4078075	Upstream	BAS4161	ArsR family transcriptional regulator	0.11	0.18
4080624–4080657	Within	BAS4164	Hypothetical protein	0.24	ND
4108375–4108387	Upstream	BAS4193	Cytochrome c-550	0.12	ND
4108498–4108548	Upstream	BAS4193	Cytochrome c-550	0.12	ND
4167156–4167184	Upstream	BAS4253 (phhA)	Phenylalanine 4-monooxygenase	0.11	0.11
4171735–4171781	Upstream	BAS4259	Acyl dehydratase MaoC	0.11	0.14
4204548–4204603	Upstream	BAS4293*	Rrf2 family protein	ND	0.49
4320800–4320819	Upstream	BAS4411	4-Hydroxybenzoyl-CoA thioesterase	0.18	0.20
4330544–4330717	Upstream	BAS4421	TetR family transcriptional regulator	0.16	ND
4332506–4332520	Upstream	BAS4422*	Long-chain-fatty-acid-CoA ligase	0.28	ND
4419232–4419254	Upstream	BAS4502	Hypothetical protein	0.19	ND
4426207–4426226	Upstream	BAS4512	Hypothetical protein	0.08	0.16
4450461–4450480	Upstream	BAS4543*	Acetyl-CoA synthetase	0.02	0.03
4466371–4466481	Upstream	BAS4560	Acetyl-CoA synthetase	0.05	0.14
	Upstream	BAS4561*	Acetoin utilization protein AcuA	0.08	0.21
4468996–4469022	Start/within	BAS4563	Acetoin utilization protein AcuC	0.07	0.17
4669518–4669533	Upstream	BAS4778*	Hypothetical protein	0.13	0.17
4830500–4830510	Within	BAS4949	Iron ABC transporter ATP-binding protein	7.1	ND
4841944–4841976	Upstream	BAS4960	Holin-like protein	ND	2.1
4956684–4956861	Upstream	BAS5077	l-Lactate permease	0.26	0.41
5075428–5075461	Within	BAS5194*	Ferredoxin, 4Fe-4S	0.10	0.34
5076900–5076944	Upstream	BAS5194*	Ferredoxin, 4Fe-4S	0.10	0.34
5097124–5097156	Upstream	BAS5208	Aminopeptidase	0.02	0.03
5139237–5139259	Upstream	BAS5256	Homoserine dehydrogenase	ND	0.43
5152423–5152450	Upstream	BAS5269	Carbon starvation protein A	0.12	0.11
5191639–5191676	Upstream	BAS5304	UDP-glucose 4-epimerase	0.10	0.10
5197185–5197217	Upstream	BAS5310	Oligoendopeptidase F	0.13	0.22
pXO1
125472–125643	Within	gbaA_pXO1_0143 (apt)	Pseudogene	3.4	3

^aCoordinates of the enriched region.

^bMost probable CodY target; two targets may be possible when the binding region is between two divergently expressed genes.

^cGenes marked with an asterisk are the first genes of a CodY-controlled operon.

^dExpression ratio (B. anthracis 7702 to codY mutant) of the CodY target gene at A₆₀₀ of 0.2 and 1.0 (11). Boldface type indicates genes that are activated by CodY.

^eND, no significant difference in expression between B. anthracis 7702 and the codY mutant; CoA, coenzyme A; HAD, haloacid dehalogenase.

Two hundred fifty-seven regions of CodY binding were not linked to genes whose expression levels are known to be controlled by CodY (11). Of these, 52 regions were outside apparent regulatory regions and not close to putatively CodY-controlled genes; 37 of these 52 regions were within a coding sequence (located >150 bp downstream of the translation start codon), and 15 of the 52 regions were located between two convergently transcribed genes. These results suggest that CodY binds in vitro to ≥52 sites at which such binding activity may not affect the transcription of any neighboring genes.

A common motif was sought in the DNA regions that bound CodY using the MEME suite (http://meme.sdsc.edu/meme4_6_1/cgi-bin/meme.cgi). As shown in Fig. 1, a common motif was found in 366 sequences with an overall E value of 1.0 × 10⁻²⁴³. The sequence with the least adherence to the motif had a P value of 3.96 × 10^−0.2. Compared to the CodY-binding motifs described for Lactococcus lactis (22, 23), B. subtilis (38), Staphylococcus aureus (25), and Clostridium difficile (24), the motif we retrieved resembles the canonical CodY box without the five nucleotides at the 5′ end.

Fig 1.

B. anthracis CodY-binding motif. A search for common motifs in the 375 chromosomal and plasmid regions enriched in the CodY pulldown experiment was conducted using the MEME suite and the parameters described in Materials and Methods. The motif was found in 366 of the 375 enriched regions, with an E value of 1.0 × 10⁻²⁴³. For reference, the canonical Bacillus CodY box identified previously (38) is shown below the B. anthracis de novo binding motifs.

B. anthracis CodY interacts with the atxA promoter region.

Our affinity purification results indicated that CodY binds to the atxA promoter region. However, our previous results showed that a codY mutation has only a small effect on atxA transcription, specifically during the exponential growth phase (11). We consequently decided to verify the binding of CodY to the atxA promoter region using electrophoretic mobility shift assays (Fig. 2A) and DNase I footprinting experiments (Fig. 2B). The apparent K_d (dissociation constant) for CodY binding to the atxA DNA sequence in the presence of GTP and branched-chain amino acids (BCAAs) was 4 to 8 nM, indicating high-affinity binding. Furthermore, in the absence of the effectors, the K_d was 128 nM; the higher binding activity in the presence of effectors is indicative of specificity. The DNase I footprinting assay revealed a CodY-binding site located between positions +38 and +76 with respect to the P1 promoter (4), which overlaps with the sequence identified in the affinity purification experiment. The apparent contradiction between the transcriptional analysis and the in vitro binding data might be a consequence of the absence of a hypothetical regulatory factor needed for high-level transcription of atxA in exponential-phase cells. Alternatively, a second repressor of atxA might be active in exponential-phase cells. In either case, the inactivation of CodY would have little or no effect on atxA expression during exponential-phase growth.

Fig 2.

Analysis of B. anthracis CodY binding to the atxA promoter region. (A) Electrophoretic mobility shift assays were carried out in the presence of increasing concentrations of CodY and in the presence or absence of CodY coeffectors: 10 mM (each) isoleucine, leucine, and valine and 2 mM GTP. (B) DNase I protection assays of CodY binding to the atxA promoter. A PCR product amplified using the primers oPJ29 and oPJ44 was subjected to treatment with DNase I after incubation with CodY. The concentrations of CodY (nM) in lanes 1 to 10 were 0, 0.5, 1, 2, 4, 8, 16, 32, 64, and 0 nM. Each binding reaction mixture contained 2 mM GTP and 10 mM (each) isoleucine, leucine, and valine. The coordinates denote the regions protected by CodY. The −10 region of the atxA promoter and the transcription start point (bent arrow) are also indicated. A sequencing ladder is shown on the left-hand side of the panel.

B. anthracis CodY interacts with the sap and eag promoter regions.

Repression by the direct binding of CodY to the sap and eag promoter regions, as suggested by the affinity purification results, would explain the overexpression of these genes in the codY mutant, as indicated by the microarray data (11). To confirm the validity of the sequence analysis, we showed by gene-specific PCR that the sap and eag regulatory regions were enriched in the copurified DNA fragments (data not shown). We then tested the ability of purified B. anthracis CodY to interact with these promoter regions using electrophoretic mobility shift assays for sap and DNase I footprinting experiments for both sap and eag (Fig. 3). High-affinity binding of CodY was detected for both promoters. For sap and eag, the apparent K_ds for CodY binding in the presence of GTP and BCAAs were 8 nM and 80 nM, respectively (Fig. 3). CodY protected the sap promoter region between positions +48 and +79 with respect to the transcriptional start point, a region that overlapped well with the binding site revealed by affinity purification (Fig. 3B) (Table 2). Since this binding site is located downstream of the transcriptional start site, CodY is very likely to repress sap transcription by acting as a roadblock to RNA polymerase (39). CodY protected two regions of the eag locus, corresponding to positions −86 to −56 and +93 to +123 with respect to the σ^H-dependent eag transcriptional start site, and positions −316 to −286 and −137 to −107 with respect to the second eag transcriptional start (the major transcriptional start site of eag when cells are grown in RBic medium [18, 19]). Both of these regions were also identified by affinity purification (Fig. 3B) (Table 2). The PagR-binding site (positions +45 to +94 with respect to the σ^H-dependent transcriptional start site) and the CodY-binding sites identified for eag are directly adjacent to each other.

Fig 3.

Analysis of B. anthracis CodY binding to the sap and eag promoter regions. (A) Electrophoretic mobility shift assays of CodY binding to the sap regulatory region were carried out with increasing concentrations of CodY and in presence or absence of effectors (see legend to Fig. 2). (B) DNase I protection assays of CodY binding to the sap (left panel) and eag (right panel) promoters. Each binding reaction mixture contained 2 mM GTP and 10 mM (each) isoleucine, leucine, and valine. The vertical lines and coordinates denote the regions protected by CodY. Sequencing ladders are shown for each of the probes. For the sap promoter, a 273-bp PCR product was amplified using primers oPJ37 and oPJ38 and was subjected to treatment with DNase I after incubation with CodY. The concentrations of CodY (nM) in lanes 1 to 10 were 0, 2, 4, 8, 16, 32, 64, 128, 256, and 0 nM. For the eag promoter, the DNA fragment was labeled at either extremity. In lanes 1 to 7, the 567-bp probe was synthesized using oPJ36 and radioactively labeled oPJ33; in lanes 8 to 14, the probe was synthesized using oPJ33 and radioactively labeled oPJ36. The CodY concentrations were 0 in lanes 1 and 8, 5 nM in lanes 2 and 9, 10 nM in lanes 3 and 10, 20 nM in lanes 4 and 11, 40 nM in lanes 5 and 12, 80 nM in lanes 6 and 13, and 160 nM in lanes 7 and 14. The positions are given with respect to the transcriptional start site (the σ^H-dependent start site for eag).

Table 2.

CodY-binding sites in the sap and eag regulatory regions^a

Target gene	Binding site by footprinting	Binding site by affinity purification	MEME-determined CodY motif within canonical CodY box
sap	+48 to +79	+21 to +83	+50 TAAAATCAGAATATTT +65
eag	−86 to −56	−80 to −53	−79 TGTAGTCTGAAAATTG −64
	+93 to +123	+90 to +119	+104 CACATTCTGATAATTC +119

^aSites of CodY binding determined by DNase I footprinting and affinity purification were associated with motifs corresponding to the MEME-determined consensus motif for B. anthracis CodY-binding sites (in boldface type, TCAGAAAATTC) and the canonical CodY box (AATTTTCNGAAAATT) determined for other bacteria. For the eag gene, the coordinates of the binding sites are relative to the σ^H-dependent transcription start site.

CodY is a negative regulator of S-layer genes sap and eag.

Previous microarray analyses showed that sap and eag transcription levels are higher in the codY mutant strain than in its parent (11). To confirm and quantitate this result, parental and codY strains carrying sap-lacZ and eag-lacZ reporter fusions were grown in RBic and sampled at multiple time points for the determination of β-galactosidase activity (Fig. 4). Whereas in the parental strain sap was barely expressed during any phase of growth, it was transcribed at high levels during the exponential phase in the codY mutant strain, indicating that this gene is normally repressed by CodY (Fig. 4A). The expression level of sap in the codY mutant dropped when cells entered the stationary phase, but the gene was still transcribed at higher levels in the mutant strain than in the parental strain. The eag gene was also transcribed at a higher level during the exponential phase in the codY mutant strain than in the parental strain, again indicating repression by CodY in the parental strain (Fig. 4B). When cells entered the stationary phase, eag expression dropped in the codY mutant strain but increased in the wild-type strain. This result implies that repression by CodY is not the only regulatory mechanism that is involved in the temporal control of eag expression.

DISCUSSION

CodY targets.

Our genome-wide analysis of in vitro DNA binding of CodY yielded 389 DNA regions that are bound by CodY. Among these, 132 regions are close to 131 transcription units that were identified previously as being regulated by CodY through expression profiling (11), while the remaining 257 regions identified by genome-wide analysis of DNA binding by CodY were not identified previously as targets of CodY-mediated regulation. Majerczyk et al. obtained similar results when they analyzed the S. aureus genome for CodY-binding sites using a similar technique (25). Microarray analysis might underestimate the effects of a regulatory protein for several reasons. For instance, CodY might control some genes whose levels of expression are below the level of detection in microarray experiments under the growth conditions used. Alternatively, the extent of regulation by CodY might be less than the 2-fold cutoff arbitrarily imposed previously, either because CodY has only a small effect on gene expression or because the effect of CodY is masked by those of other regulatory proteins. For instance, atxA was not identified previously as a CodY target by transcription profiling, and lacZ transcriptional fusion data indicated that a codY mutation has only a modest effect on atxA transcription and only during the exponential growth phase (11). Nonetheless, the atxA regulatory region was copurified with CodY, and gel mobility shift and DNase I footprinting assays confirmed this binding activity. The lack of effects of a codY mutation on transcription may be explained by the activity of AbrB, a repressor of atxA transcription during the exponential growth phase (10, 40, 41). AbrB binds to the atxA promoter region upstream of the transcription initiation site. CodY, on the other hand, binds between the transcription and translation start sites; thus, the binding sites of AbrB and CodY do not overlap. An alternative explanation for the minimal effect of a codY mutation on atxA transcription might be that an unidentified positive regulator is needed for significant atxA expression and that this regulator is active only in stationary-phase cells. These explanations for why some CodY-binding sites have not been tied to transcriptional regulation notwithstanding, one needs to consider the possibility that some CodY-binding sites are nonfunctional or have functions unrelated to transcription regulation. In some cases, the CodY-binding sites are located within the coding sequences of genes that do not appear to be under CodY control or between two convergently transcribed genes. It has been suggested that Escherichia coli RutR recognizes numerous nonfunctional binding sites that have not been eliminated by evolution because binding to those sites has no adverse effects (42). It is possible that similar unevolved sites were also identified in our CodY-DNA binding assay.

Applying the MEME algorithm to all 375 affinity-purified sites of ≥8 bp, we found an 11-nucleotide nonpalindromic consensus motif. While this 11-bp sequence might be important for CodY binding, it is unlikely to be sufficient, since it is found with fewer than 2 mismatches at more than 1,600 positions in the B. anthracis genome, i.e., at many more sites than those copurified with CodY. Since perfect palindromes might be unstable in bacteria (43, 44), this sequence might be part of a palindrome in which the nucleotides that are not absolutely required for CodY binding are allowed to vary. Their occurrence might be below the threshold recognized by MEME. It is notable that the CodY-DNA binding motif found here shares 10 of the 15 bases in the consensus sequence described for other bacteria (22–25, 38, 45, 46). This result is not surprising, since the 26-residue helix-turn-helix motifs of B. anthracis CodY and B. subtilis CodY are identical and differ from the S. aureus motif at a single position that is not thought to interact with DNA. On the other hand, a search of the B. anthracis genome for sequences matching the canonical CodY box (AATTTTCWGAAAATT) revealed 1 site with no mismatches, 11 sites with 1 mismatch, and 91 sites with 2 mismatches. There is only a partial correlation between the sequences matching this consensus sequence and those copurified with CodY, however, and the relationship between the motif derived for B. anthracis and the affinity of CodY binding remains unclear.

CodY bound to 131 promoter regions, thus directly controlling 197 genes of the 500 previously shown to be CodY-controlled by microarray analysis. Among the direct targets of CodY are the genes that code for the biosynthesis of the CodY effectors, GTP and the branched-chain amino acids (BCAAs). The GTP biosynthesis gene BAS0011 is positively regulated by CodY, whereas the operons for BCAA biosynthesis (with the promoter-proximal genes BAS1307, BAS1312, and BAS1713) are repressed by CodY. The breadth of metabolic genes regulated directly by CodY in B. anthracis is very reminiscent of those under CodY control in B. subtilis, S. aureus, and C. difficile (11, 24, 25, 47). Among the B. anthracis genes that are indirectly controlled are those involved in iron uptake and detoxification and most of the CodY-controlled genes located on pXO1. We have shown that the three toxin genes carried by pXO1 are indirectly controlled through CodY-dependent accumulation of AtxA (11). B. anthracis is the first example of a bacterium for which CodY activity is needed for virulence (11, 12), due to its indirect control of the expression of its toxin and iron uptake genes and detoxification genes. A recent report showed that Listeria monocytogenes CodY is needed for the expression of the virulence regulatory gene prfA and is therefore also required for virulence (48).

The observation that CodY directly controls the expression of nine genes encoding transcriptional regulators (such as NprR, SinR, and the sigma factor σ^H) probably accounts for its indirect control of many of the remaining genes. For example, NprR controls the expression of nprA in Bacillus cereus, a close relative of B. anthracis, and nprA was found to be repressed 7-fold in our microarray analysis (11, 49); SinR is a pleiotropic regulator in B. subtilis and in B. anthracis (50, 51); and σ^H, as an RNA polymerase cofactor, presumably controls a substantial regulon, which, in B. subtilis, consists of 87 genes (52). The exact role of σ^H in B. anthracis is uncertain. Involvement in the regulation of the atxA gene has been proposed (53), but that conclusion is controversial (54). Only two B. anthracis homologs of genes dependent on σ^H for their transcription in B. subtilis (53) were among the genes shown to be under CodY control (spoVG in BA0047 and citG in BA1767) (11). They are repressed and activated, respectively, by CodY during the exponential phase. Since neither of these genes was associated with a CodY-binding site in the present study, spoVG might be regulated indirectly by CodY through σ^H, whereas activation of the expression of citG in a codY mutant might reflect the control of citG by an unidentified regulatory protein. Interestingly, eag, which has a σ^H-dependent promoter, is under direct CodY control, raising the possibility that CodY regulates eag in two ways, by directly and indirectly controlling σ^H synthesis.

CodY directly represses expression of S-layer genes.

The microarray results indicated that expression levels of the S-layer genes sap and eag are derepressed in the absence of CodY (11). The data presented here show that CodY represses the expression levels of both of these genes during the exponential phase. Furthermore, affinity purification and in vitro assays of CodY binding to the sap and eag promoter regions indicate that CodY acts directly on these genes. From the early stationary phase onwards, the regulation of the expression of sap and eag becomes more complex. In the parental strain, the regulator PagR, which is encoded by a small open reading frame that lies in an operon with pagA, is responsible for the downregulation of sap transcription and the activation of eag expression (18, 20). In a codY mutant, the pagAR operon is not as highly expressed as it is in the parental strain (11) and the reduced levels of PagR that result from this might explain the limited repression of sap transcription and the limited activation of eag that are seen when the codY mutant reaches the stationary phase. The intervention of two regulators, one responding to bicarbonate and one to nutritional and energy cues, suggests that the expression of the S-layer genes is tightly controlled and depends on diverse cues.

Our data further show the extraordinary complexity of the mechanisms used by B. anthracis to regulate the expression of genes under conditions that mimic those prevalent in the mammalian host. CodY has an important role in these processes, but additional regulatory pathways exist, many of which remain to be elucidated.