• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Apr 1, 2003; 100(7): 3808–3813.
Published online Mar 24, 2003. doi:  10.1073/pnas.0737977100
PMCID: PMC153003
Bioinorganic Chemistry Special Feature

Metal stoichiometry and functional studies of the diphtheria toxin repressor


Diphtheria toxin repressor (DtxR) is a transition metal ion-activated repressor in Corynebacterium diphtheriae. DtxR is an iron sensor; metal-bound DtxR represses transcription of genes downstream of the tox operator. Wild-type DtxR [DtxR(wt)] and several mutant forms were overexpressed and purified from Escherichia coli. DtxR was isolated without bound metal. Metal reconstitution gave a binding stoichiometry of 2 per monomer for DtxR(wt) and 1 per monomer for DtxR(H79A) and DtxR(M10A). DNA binding of DtxR(H79A) and DtxR(M10A) indicates that metal site 2 is essential for activity. Metal binding lowers the dimerization Kd of DtxR from low micromolar to 33 nM. Gel electrophoretic mobility-shift assays show that Fe2+ and not Fe3+ activates DtxR for DNA binding. This finding suggests that gene regulation by DtxR may be sensitive not only to iron levels but also to redox state of the iron. Mutations in the tox operator sequence indicate that DtxR dimers binding to DNA may be highly cooperative.

Production of diphtheria toxin by toxinogenic isolates of Corynebacterium diphtheriae is affected dramatically by the culture medium. Early studies demonstrated that diphtheria toxin production is greatest when C. diphtheriae is grown under iron-limiting conditions and is severely inhibited with high concentrations of iron (1). Biochemical and genetic evidence has shown that diphtheria toxin is synthesized by C. diphtheriae lysogenic for one of a family of corynebacteriophages that carries the gene for the toxin (tox). A corynebacterium-derived iron-binding repressor, DtxR, regulates the corynebacteriophage tox gene. Production of diphtheria toxin and the siderophore corynebactin is coordinately regulated during the transition from iron-replete to iron-deficient growth (2). In addition, heme oxygenase in C. diphtheriae is regulated by DtxR, and five different iron-regulated operators repressible by DtxR have been identified (37). Thus, DtxR serves as a global regulatory protein to coordinately regulate the expression of diphtheria toxin, the high-affinity iron uptake system, and other iron-sensitive genes of C. diphtheriae in response to environmental iron concentrations.

DtxR is a 226-amino acid polypeptide, which functions as a homodimer (811). Recombinant DtxR was tested in vitro for interaction with DNA containing the tox operator region by gel electrophoretic mobility-shift assays (EMSAs) and DNase I and hydroxyl radical footprinting (1215). These studies demonstrated that DtxR binding to the tox operator requires divalent cations and that a specific sequence of ≈30 bp is protected by DtxR. The consensus binding sequence, determined by in vitro affinity selection, is an interrupted palindrome of 19 bp that includes two overlapping imperfect palindromes with conserved bases at symmetrical positions (16).

The structures of wild-type apo- and holo-DtxR(wt) and mutant holo-DtxR(C102D) have been determined by x-ray crystallography (1722) and NMR (23, 24). Each DtxR monomer has three domains. Domain 1 (residues 1–73) contains a helix–turn–helix motif that is the DNA-binding site. Domain 2 (residues 77–144) contains the dimerization domain and the two metal-binding sites (Fig. (Fig.1).1). Metal site 1 (the ancillary site) is composed of ligands H79, E83, H98, E170, and Q173. Metal site 2 (the primary site) consists of M10, C102, E105, and H106, and the main-chain carbonyl oxygen of C102. Domain 3 (residues 145–226) exhibits an SH3-like fold, the function of which is unknown. Two crystal structures of Ni2+-DtxR(C102D) and Co2+-DtxR(wt) complexed with DNA have been solved (25, 26). Both show that two dimeric repressor proteins bind to opposite sides of the DNA major groove and do not appear to interact with one another.

Figure 1
The metal-binding sites of Co2+-DtxR(wt). The ligands to each metal on the monomer colored in blue are indicated, including both the amide carbonyl and thiol of C102 (site 1 is the upper site and site 2 the lower). Also shown are the tryptophans ...

The expression and purification of DtxR(wt) and two mutant forms of DtxR plus a metal reconstitution procedure for DtxR that allows for the determination of metal stoichiometry under physiological conditions and relative metal-binding affinities are reported as well as other properties of this protein.

Experimental Procedures

Construction of DtxR(M10A) Expression Vector.

A DtxR(wt) construct in pET11c (pDR1) was previously described (13). The DtxR(M10A) expression vector was made by PCR using a mutagenic sense primer containing the ATG start codon in a unique NdeI site and the two underlined base changes necessary for the M10A mutation (5′-CATATGAAGGACTTAGTCCATACCACAGAGGCGTATTG-3′). The antisense primer was positioned outside of the DtxR gene and contained a unique AgeI site (5′-CCACCGGTAACGAGGAGTTT-3′). PCR of pDR1 yielded a 729-bp fragment, which was cloned into pETBlue-1 Blunt vector by using a Perfectly Blunt Cloning kit (Novagen) and transformed into NovaBlue competent cells (Novagen). Recombinant colonies were screened by colony PCR. Plasmid pETBlue-1 with the DtxR(M10A) insert was digested with NdeI and AgeI and ligated into similarly digested pDR1 to make pET11c-DtxR(M10A). Purified plasmid DNA was sequenced by the University of Michigan Biomedical Research Core Facilities.

Expression and Purification of DtxR.

The following procedure was used to express and purify both wild type and mutant forms of DtxR. E. coli HMS174(DE3) transformed with the appropriate plasmid [pDR1 for wild type (13), pET11c-DtxR(H79A) for the H79A mutant (19), and pET11c-DtxR(M10A) for the M10A mutant] were grown at 37°C in Miller's Luria broth supplemented with 100 μg/ml ampicillin. Expression was induced at OD600 of ≈0.6 with 0.4 mM isopropyl β-d-thiogalactoside. Two hours after induction, bacteria were centrifuged and frozen at −80°C. All purification steps were carried out at 4°C. Cell pellets from 6 liters of culture were suspended in 50 ml of 10 mM Tris[center dot]HCl (pH 7.5)/5 mM 2-mercaptoethanol, lysed in a French pressure cell (SLM-Aminco, Rochester, NY), and centrifuged for 45 min at 30,000 × g. Supernatant was loaded at 1 ml/min onto a 30-ml DEAE Bio-Gel A anion-exchange column equilibrated with 10 mM Tris[center dot]HCl (pH 7.5), then washed with 150 ml of 10 mM Tris[center dot]HCl (pH 7.5), and DtxR was eluted with a 300-ml linear gradient of 0–300 mM NaCl in 10 mM Tris[center dot]HCl (pH 7.5). Fractions containing DtxR were identified by SDS/PAGE. Pooled fractions were applied to a 10-ml Ni-nitrilotriacetate (NTA)-agarose column equilibrated with 10 mM Tris[center dot]HCl (pH 7.0). The column was washed with 40 ml 10 mM Tris[center dot]HCl (pH 7.0), and a 100-ml linear gradient of 0–10 mM imidazole in 10 mM Tris[center dot]HCl (pH 7.0) was used to elute the DtxR. Fractions containing DtxR were concentrated to <1.5 ml by using Ultrafree-15 concentrators (nominal molecular weight limit 10,000; Millipore), and 10 mM DTT was added. The concentrate was applied to a Superdex 200 HiLoad 16/60 (Amersham Pharmacia–Pharmacia Biotech) gel filtration column equilibrated with 10 mM Tris[center dot]HCl (pH 7.5)/100 mM NaCl/10 mM DTT. Fractions containing DtxR were concentrated and frozen at −80°C with fresh DTT (10 mM). N-terminal sequences were determined by the University of Michigan Biomedical Research Core Facilities.

Protein Quantitation.

Protein quantitation was by the Bradford method (27) with a BSA standard. The Bradford assay was standardized against quantitative amino acid analysis (QAA) and dry weight measurement of DtxR. QAA was carried out by the University of Michigan Biomedical Research Core Facilities on samples containing ≈1 mg/ml DtxR(wt) in 10 mM Tris[center dot]HCl (pH 7.5). Sample concentration was determined by the Bradford method before and after analysis; in all cases the concentrations before and after QAA were in good agreement. The alanine content of the samples was used to determine the protein concentration. The dry weight of DtxR was determined by desalting samples of DtxR (7.5 mg) into water by using a PD-10 column (Amersham–Pharmacia Biotech). Aliquots of desalted protein (≈6 mg of DtxR) were then dried in a 104°C oven until a constant weight (≈6 h). Parallel samples were assayed by the Bradford method.

Preparation of Metal-Free Labware and Buffers.

Clear plastic tips and microcentrifuge tubes were used because colored plastics are often tinted with metals. Glassware was washed with 10% nitric acid and rinsed with metal-free water. Metal-free buffers were made by batch-chelating buffers overnight with Amberlite IRC748 resin (Supelco). The pH was then adjusted, and the buffer and resin were stirred slowly for another 2–3 h. The buffer was brought to its final volume with metal-free water after the resin had been removed with a disposable filter.

Metal Reconstitution of DtxR.

Apo-DtxR (275 μl of 50 μM to 1 mM) was exchanged into metal-free 10 mM Tris[center dot]HCl (pH 7.5) by using a PD-10 column and, if necessary, concentrated by using Ultrafree-0.5 (nominal molecular weight limit 10,000) concentrators. DtxR was then incubated at room temperature for 30–60 min with 5 mM metal [CoCl2·6H2O, NiSO4·6H2O, MnCl2·6H2O, FeCl3·6H2O, or Fe(NH4)2(SO4)2·6H2O]. Unbound metal was removed by using a PD-10 column equilibrated with metal-free 10 mM Tris[center dot]HCl (pH 7.5). Reconstitutions with Fe3+ were performed with sodium citrate (1:10 Fe/sodium citrate) and protected from light to prevent photoreduction of the iron.

Reconstitutions with Fe2+ were performed as above except that all steps were carried out in an anaerobic chamber (Coy Laboratory Products, Grass Lake, MI). Apo-DtxR (275 μl of 50 μM to 1 mM) was made anaerobic in vials with Teflon/silicone septa on a gas train by using 10 vacuum/argon cycles. Solutions of Fe2+ were made in the anaerobic chamber by dissolving preweighed dry ferrous ammonium sulfate in deoxygenated metal-free water; a few grains of sodium dithionite were added to reduce any Fe3+.

Metal Analysis.

Metal analysis was carried out by using a Finnigan Element inductively coupled plasma high-resolution mass spectrometer (ICP–MS). The metal content of DtxR samples exchanged into metal-free 10 mM Tris[center dot]HCl (pH 7.5), reconstituted DtxR samples, and buffer controls were determined.

Iron quantitation was also performed by a colorimetric method (28), with the following modifications. Protein samples (up to 100 μl) were boiled for 30 min in a total volume of 200 μl of 2 M HCl. Precipitated protein was pelleted by centrifugation, and freshly prepared sodium ascorbate (40 μl of 75 mM) and 200 μl of 10 mM ferrozine [3-(2-pyridyl)-5,6-bis(4-phenylsulfonic acid)-1,2,4-triazine] were added to the supernatant. The Fe2+(ferrozine)3 complex became magenta with the addition of 200 μl of a saturated ammonium acetate solution. The A562 was then measured (epsilon562 = 27,900 M−1[center dot]cm−1) (29). When measuring Fe2+ in reconstituted DtxR, 1 mM (final concentration) ferrozine was added directly and the A562 was measured without boiling or additions.


Far-UV CD spectra were taken on an Aviv 62DS spectropolarimeter at 25°C. Apo-DtxR samples (≈0.25 mg/ml) were exchanged into metal-free 10 mM Tris[center dot]H2SO4 (pH 7.5) by using a PD-10 column. Holo-DtxR samples (≈0.25 mg/ml) were prepared identically except 200 μM NiSO4 was added after buffer exchange. Spectra were recorded from 190 to 240 nm with a data interval of 0.5 nm, an integration time of 15 s, and a pathlength of 1 mm. After baseline correction, the spectra were normalized to the protein concentration of the sample.

Analytical Ultracentrifugation.

Sedimentation equilibrium data were collected on a Beckman XL-A analytical ultracentrifuge at 20°C and 12,000, 17,000, and 24,500 rpm for apo-DtxR(wt) and additionally at 32,000 rpm for Co2+-DtxR(wt). Samples were prepared by exchanging DtxR(wt) (300 μl of 5 mg/ml) into 10 mM Tris[center dot]HCl (pH 7.5)/150 mM NaCl/0.005% Surfactant P20 and either 10 μM EDTA or 1 mM CoCl2 by using a PD-10 column. Protein concentration ranged from 0.4 to 40 μM. Samples were monitored at 280 nm for concentrations ≥4 μM and at 220 nm for concentrations <4 μM. All experiments used six-channel, 12-mm-thick charcoal–Epon centerpieces and matched quartz windows in wide-aperture window holders. These data were fit by nonlinear regression analysis by using the program winnonln (National Analytical Centrifugation Facility). The simplest equation to which the data fit was a monomer–dimer equilibrium model. The dimerization Kd was calculated by using epsilon220 = 3.1 × 105 M−1[center dot]1.2 cm−1.

Surface Plasmon Resonance (SPR).

SPR analysis was carried out on a Biacore 3000 system at 25°C. Biotinylated DNA hairpins (Table (Table1)1) were immobilized on a Sensor Chip SA; the unmodified surface was used as a reference. The continuous-flow buffer was metal-free 10 mM Tris[center dot]HCl (pH 7.5)/150 mM NaCl/0.005% Surfactant P20 at a flow rate of 20–100 μl/min. Apo-DtxR (300 μl of 33–165 μM) was made anaerobic as described above and exchanged into deoxygenated, metal-free continuous-flow buffer by using a PD-10 column. DNA binding of DtxR was tested on surfaces with 120–710 resonance units (RU) of immobilized DNA and 280–420 nM DtxR in continuous flow buffer with the specified divalent metal concentration. For DNA-binding Kd determination, surfaces with 11–26 RU of immobilized DNA and 0–2 μM DtxR(wt) in continuous-flow buffer with 20 μM Co2+ were used. The surface was regenerated after each analyte injection with 10 μl of 0.01% SDS/3 mM EDTA. CLAMP99 (30) was used to analyze data, using a model with two independent classes of binding sites.

Table 1
Sequences of 5′-biotinylated DNA hairpins


Binding of DtxR(wt) to DNA probes was carried out in 16 μl of reaction mixture containing 100 mM Tris[center dot]HCl (pH 7.5), 5 mM MgCl2, 40 mM KCl, 10% (vol/vol) glycerol, 1.5 μg of apo- or reconstituted DtxR, and 36 ng of DNA probe. Reaction mixtures with DNA hairpins also contained 2 mM DTT and 125 μM MnCl2 with Mn2+ present in the gel and running buffer. The toxPO probe was amplified from pRS551toxPO (8) by PCR using primers 5′-GCAGAATTCTGCAGGGCATTGA-3′ and 5′-CATGGATCCAGGACTCATAA-3′. After 10 min at 25°C, 2 μl of loading buffer (bromophenol blue in reaction buffer) was added. The entire mixture was applied to a 6% polyacrylamide gel in 40 mM bis(2-hydroxyethyl)iminotris(hydroxymethyl)methane (Bis-Tris) borate (pH 7.5)/2.5% glycerol and electrophoresed in the same buffer without glycerol. Gels were stained with Sybr green I nucleic acid gel stain (Molecular Probes) and visualized by using a Fluor-S MultiImager (Bio-Rad) with UV excitation and a 530-nm bandpass filter.

EMSAs with Fe2+-DtxR(wt) were carried out anaerobically. The polyacrylamide gel was poured in the anaerobic chamber. For EMSAs with Fe3+-DtxR(wt), apo-DtxR(wt) was reconstituted with Fe2+ and subsequently oxidized by diluting with oxygen-saturated buffer and incubating under an atmosphere of 100% oxygen. Ferrozine was used to measure the amount of remaining Fe2+ as described above. The oxidation of Fe2+, reaction incubations, and electrophoresis were done in the dark to prevent the photoreduction of Fe3+.

Quantitation of Free Sulfhydryls.

The free sulfhydryl content of DtxR(wt) was determined with DTNB [5,5′-dithiobis(2-nitrobenzoic acid)] and using a cysteine standard (31).


Expression and Purification of DtxR.

Modifications made to the published methods (13, 32) simplified the procedure and increased protein yield. The most significant change was replacement of the DEAE-Sepharose column with a gel filtration column and using DTT to reduce disulfide-linked dimeric DtxR, leading to a higher yield of monomeric DtxR.

Isopropyl β-d-thiogalactoside induction of DtxR from a T7 promoter in E. coli HMS174(DE3) resulted in the overproduction of a protein that migrated at a position corresponding to 28 kDa by SDS/PAGE (predicted molecular mass is 25.3 kDa). DtxR was purified to >95% purity by using anion-exchange, metal-affinity, and gel-filtration chromatography as judged by SDS/PAGE (data not shown). N-terminal sequence analysis confirmed the identity of the purified protein (data not shown). A typical purification of DtxR(wt) or DtxR(M10A) yielded 60–70 mg of protein from 6 liters of cell culture. DtxR(H79A) purification yields were slightly less, resulting in 40–50 mg of protein, most likely attributable to a loss of protein on the Ni-NTA column.

Protein Quantitation.

Determination of the DtxR concentration was critical for establishing metal stoichiometries; therefore standardization of the Bradford method was required. Bradford correction factors determined from QAA and dry weight measurements of DtxR agreed very closely; a correction factor of 0.8 ± 0.1 was determined from both QAA (n = 5) and dry weight measurement (n = 8). The Bradford method with a BSA standard overestimates the actual DtxR concentration.

Metal Content of Reconstituted DtxR.

The metal content of DtxR as isolated and after reconstitution was measured (Table (Table2).2). After purification, DtxR contained no bound metal. DtxR(wt) reconstituted with Co2+, Ni2+, or Fe2+ bound up to 2 eq of these metals per monomer; Mn2+ bound to a lesser extent (1.3 ± 0.2 per monomer). In contrast, DtxR(H79A) and DtxR(M10A) reconstituted with Co2+, Ni2+, or Fe2+ bound at most 1 eq of these metals per monomer. Reconstitutions with Mn2+ resulted in 0.21 ± 0.02 Mn2+ ion bound per DtxR(M10A) monomer, and no significant amount of Mn2+ bound to DtxR(H79A). Metal stoichiometries were significantly higher for DtxR proteins reconstituted with Fe3+. Ferric iron forms an insoluble iron hydroxide at physiological pH; high Fe3+ stoichiometry probably results from nonspecific binding of iron hydroxide to the protein surface.

Table 2
Metal content of DtxR as isolated and after reconstitution

CD of DtxR.

Far-UV CD spectra of DtxR mutants with and without Ni2+ bound were compared with spectra of DtxR(wt) with and without Ni2+ bound (data not shown). CD spectra in the absence of metal are very similar for all of the DtxR proteins; reconstitution with Ni2+ did not produce any significant changes in the CD spectra. This result is consistent with crystal structure data, where only minor structural changes were observed between structures of apo- and holo-DtxR (18, 22).

Dimerization Kd.

Table Table33 shows the apparent molecular weights determined for apo-DtxR. DtxR displayed nonideal behavior at protein concentrations >4 μM, indicated by the decrease in the apparent molecular weight. There appears to be metal contamination from the charcoal–Epon centerpieces as well (despite added EDTA), because the apparent molecular weight is not dependent on the protein concentration <4 μM DtxR. Even though the data could not be fit to a monomer–dimer equilibrium model they suggest that the dimerization Kd of apo-DtxR(wt) is in the low micromolar range. Concentration gradients and global fits using a monomer–dimer equilibrium model for Co2+-DtxR(wt) are shown in Fig. Fig.2.2. Analysis of samples with high protein concentrations showed evidence of higher-order aggregates and these samples were not used in the fit. Binding of divalent metal promotes DtxR dimerization, lowering the Kd to 33 nM.

Table 3
Apparent molecular weights of apo-DtxR determined by analytical ultracentrifugation
Figure 2
Equilibrium sedimentation of Co2+-DtxR(wt). DtxR(wt) (1.78 μM) in the presence of 1 mM CoCl2 at 12,000 (○), 17,000 (□), 24,500 ([open diamond]), and 32,000 ([open triangle]) rpm. A monomer–dimer equilibrium model was used ...

DNA-Binding Activity of DtxR.

Controls to test the binding specificity of the DtxR–wt DNA hairpin interaction were performed (data not shown). DtxR(wt) (420 nM) in 80 μM Co2+ demonstrated DNA binding that could be abolished with the addition of 600 μM EDTA. DNA binding of DtxR(wt) was also dramatically reduced by the addition of 2.0 μg of nonbiotinylated wild-type DNA hairpin. A nonspecific DNA sequence [poly(dI-dC)] had no effect on DtxR(wt) DNA binding. Therefore, the interaction between DtxR in solution and the DNA hairpin immobilized on the chip is specific.

Metal dependence of DNA binding was tested (Fig. (Fig.3).3). No DNA binding was observed in samples with apo-DtxR or Fe3+. High levels of DNA binding were observed with DtxR(wt) and DtxR(H79A) in the presence of 80 μM Fe2+, Co2+, or Ni2+. DNA binding was also observed in the presence of Mn2+; however, 500 μM Mn2+ was required to achieve the same level of DNA binding with DtxR(wt) and was still not enough with DtxR(H79A). The DNA binding of DtxR(M10A) was dramatically different. Very little DNA binding was observed in samples with Fe2+, Co2+, Ni2+, or Mn2+, even at higher metal concentrations. Of these metals, Fe2+ showed the best ability to activate DtxR(M10A). Addition of 300 μM EDTA (1.2 mM EDTA in samples containing Mn2+) abolished all DNA binding (data not shown).

Figure 3
DNA-binding activity of DtxR. Sensorgrams are shown indicating the DNA binding of 420 nM DtxR(wt) (A), 280 nM DtxR(H79A) (B), and 360 nM DtxR(M10A) (C) without metal (apo) and with Fe2+, Fe3+, Co2+, Ni2+, or Mn2+ ...

Iron Activation of DtxR.

DtxR(wt) was reconstituted with Fe2+ (1.7 Fe per monomer, 94% Fe2+, 1% oxidized sulfhydryls). The DNA binding of Fe2+-DtxR(wt) is shown in Fig. Fig.44A. Despite the fact that these assays were performed inside an anaerobic chamber, some dissociation of the protein–DNA complex was seen. Presumably this was caused by oxidation of iron by oxygen generated during electrophoresis. The addition of 2,2′-dipyridyl abolished the DNA binding of Fe2+-DtxR(wt). Protocatechuic acid, a Fe3+ chelator, had minimal effect on DNA binding; however, ferrozine, a Fe2+ chelator, inhibited binding. EMSAs with Fe3+-DtxR(wt) were carried out with Fe2+-DtxR(wt) that had been oxidized by oxygen (1.7 Fe per monomer, 4% Fe2+, 15% oxidized sulfhydryls). As shown in Fig. Fig.44B, Fe3+-DtxR(wt) did not bind to DNA.

Figure 4
Iron activation of DtxR. EMSAs show the DNA binding of Fe2+-DtxR(wt) (A) and Fe3+-DtxR(wt) (B) with and without iron chelators. Lanes 1, free toxPO probe; lanes 2–5, free toxPO probe and 5 μM Fe-DtxR(wt); lanes 3, plus ...

DNA Dependence of Binding.

EMSAs and SPR were used to test the ability of wild-type and several mutated DNA sequences to bind DtxR. DNA hairpins containing the wild-type (tox operator) sequence and the palindromic consensus sequence were able to bind holo-DtxR, exhibiting a single large shift in EMSAs (data not shown). Large shifts upon DtxR binding were also seen with DNA hairpins AB2 and CD2; each hairpin contains two uracils replacing two thymine bases. However, DNA hairpins AB1, CD1, AB3, and CD3 demonstrated no evidence for binding either in EMSAs or by SPR. The stoichiometry of DtxR dimer:DNA hairpin was determined by SPR for those sequences that bound DtxR; in each case the stoichiometry was two DtxR dimers per DNA hairpin. DNA-binding constants determined from SPR kinetic analyses are given in Table Table11.


Crystal structures of DtxR revealed the possibility of two metal-binding sites per monomer (1722, 25). However, lack of full occupancy even in the presence of excess metal has left the physiologically relevant metal stoichiometry unresolved. Metal binding site 1 (ancillary site) was fully occupied in all structures, whereas occupancy at metal site 2 (primary site) has been less consistent. Only recently has full occupancy at site 2 been shown with Co2+ (26). This variability in occupancy may be due to a lower affinity for the metal at site 2 compared with site 1 (see discussion on metal affinities). It is also possible that oxidation or modification of C102 observed in several structures has interfered with metal binding. Crystal structures of a biologically active C102D mutant have also been solved to avoid the problem of C102 oxidation; however, this mutation may alter or eliminate a key metal–ligand interaction in site 2.

To accurately determine the metal stoichiometry of DtxR, a reconstitution method was developed for apo-DtxR, followed by ICP–MS metal analysis. The reconstitution procedure is quick, so oxidation of C102 is minimal, even in the absence of DTT, also eliminating the complication of DTT reducing Ni2+ and Co2+. The procedure can be carried out in an anaerobic chamber or protected from light to prevent changes in the oxidation state of iron. When the reconstitution method was used, DtxR(wt) bound up to 2 eq of Mn2+, Co2+, Ni2+, or Fe2+ per monomer. Residues H79 and M10 are important ligands for metal binding in site 1 and site 2, respectively. The metal stoichiometries of DtxR(H79A) and DtxR(M10A) were reduced to one metal ion per monomer. To our knowledge, this is the first time that DtxR(wt) has been observed to bind two full equivalents of Fe2+, the presumed physiological metal.

Holmes and coworkers (32) and Murphy and coworkers (33) have independently determined a Kd in the low to submicromolar range for Ni2+ binding to DtxR by equilibrium dialysis using 63Ni2+. However, both studies assumed only one metal-binding site per DtxR monomer. We tried several different methods to determine metal-binding Kd values, including tryptophan fluorescence and isothermal titration calorimetry. The results were inconclusive. Metal-binding titrations were also tried by using SPR to determine a Kactivity value for metal binding. Unfortunately, the inability to deconvolute metal binding and dimerization prevented further analysis.

However, relative binding affinities could be assigned from the metal stoichiometry data. DtxR binds Fe2+ and Ni2+ with about the same affinity, Co2+ with slightly less affinity, and Mn2+ with much less affinity (Fe2+ ≈ Ni2+ > Co2+ [dbl greater-than sign] Mn2+). This order is consistent with previous data showing that Mn2+ is less efficient than other metals in DNA footprinting assays (14). Metal-binding site 1 has a higher affinity than does site 2. DtxR(H79A), which has a disrupted site 1, bound fewer metal ions per monomer than DtxR(M10A), which has an intact site 1. There appears to be some cooperativity in metal binding between sites or at least some increased stability in having two occupied metal-binding sites in DtxR. DtxR(wt) bound more metal ions per monomer than DtxR(H79A) and DtxR(M10A) did combined. The CD spectra showed that the mutations do not induce gross structural changes.

There has been some debate about the role played by each of the metal-binding sites in DtxR activation. Site-directed mutagenesis was used to mutate ligands in both metal-binding sites to alanine (19). A β-galactosidase reporter system demonstrated wild-type repression for DtxR mutants H79A, E83A, and H98A, all mutations in site 1, whereas mutation of ligands in site 2, M10A, E105A, and H106A, showed a complete loss in the ability to repress β-galactosidase expression, suggesting this was the primary site required for functional repression. A comparison of the apo- and holo-DtxR crystal structures reveals that the six N-terminal residues undergo a helix-to-coil transition induced by metal binding at site 2 that relieves an unfavorable steric interaction between apo-DtxR and the DNA (25). Holmes and co-workers (20, 34) have suggested that phosphate or sulfate may act as a corepressor with the metal at site 1 to activate DtxR. A more recent crystal structure of DtxR shows that the anion-binding site has been replaced by E170 from domain 3 of the protein that until now was disordered (26). Holmes maintains that metal-binding site 1 communicates with the DNA-binding domain through a salt bridge with E170 and is therefore important for DNA binding.

This paper describes the purification and characterization of DtxR(H79A) and DtxR(M10A). The metal-binding stoichiometry has been definitively determined; each mutant binds one metal ion per monomer, presumably in the unaltered metal-binding site. In direct DNA-binding assays, DtxR(H79A) retains DNA-binding activity whereas DtxR(M10A) does not. The data indicate that metal-binding site 2 is the primary site responsible for DNA-binding activity and that site 1 is the ancillary site.

Ferrous iron is the presumed physiologically relevant metal involved in gene regulation by DtxR. Previous EMSA and DNA footprinting studies have been carried out aerobically with excess Fe2+ (12, 14) and have demonstrated DNA binding and DNA protection by DtxR; however, the oxidation state of the activating iron is uncertain. We have paid particular attention to ensuring the oxidation state of the iron in these studies. Analysis of DNA binding by SPR suggests that only Fe2+ is able to activate DtxR; DNA-binding assays with Fe3+ showed no DNA binding. However, because DtxR can bind upwards of 30 Fe3+ ions per monomer, DtxR(wt) was first reconstituted with Fe2+ and then the iron was oxidized so that the DNA-binding activity of Fe3+-DtxR(wt) could be measured without excess Fe3+ bound to the protein.

EMSA results in this paper show conclusively that Fe2+ and not Fe3+ activates DtxR. The use of iron chelators, specific for either Fe2+ or Fe3+, confirmed that the oxidation state of the iron in activated DtxR was Fe2+. The fact that only Fe2+ and not Fe3+ activates DtxR suggests that gene regulation by DtxR not only depends on iron levels but also may be redox sensitive. In particular, reactive oxygen species and/or nitric oxide (NO), which can oxidize Fe2+, may affect binding of DtxR to its target DNA sequence.

The ability of one or two DtxR dimers to serve as a functional repressor was tested. Mutations in the tox operator sequence, based on crystal structure contacts between specific DNA bases and protein residues (35), were designed to knock out binding of one DtxR dimer. Surprisingly, these mutated DNA hairpins (AB1, CD1, AB3, and CD3) showed no DNA binding in EMSAs or by SPR. DNA hairpins AB2 and CD2, with the very subtle mutation of thymine to uracil, still bound two DtxR dimers; however, the Kd of each dimer increased significantly. These results suggest that the binding of DtxR dimers to DNA is highly cooperative and that most likely two dimers are needed for repression. However, kinetic analysis of data from DNA hairpins that bound DtxR was best when a model with two classes of independent binding sites was used; this observation may be due to a limitation of the software to deal with cooperative binding models. This apparent contradiction requires more investigation.

In summary, metal reconstitution and analysis has demonstrated that DtxR binds up to two metal ions of Fe2+ under physiological conditions. DtxR binds significantly more Fe3+ because of nonspecific binding. The relative binding affinities for several divalent metals are as follows: Fe2+ ≈ Ni2+ > Co2+ [dbl greater-than sign] Mn2+. In addition, metal-binding site 1 appears to have a higher affinity for metal than site 2. The DNA-binding activities of DtxR(H79A) and DtxR(M10A) have shown the importance of metal-binding site 2 in DtxR activation; site 1 appears to be an ancillary or structural metal-binding site. The importance of the iron oxidation state for DtxR activation was studied by using Fe2+-DtxR and Fe3+-DtxR in EMSAs. The fact that only Fe2+ activates DtxR suggests that gene regulation by DtxR not only may depend on iron levels but also may be redox sensitive to reactive oxygen species and/or NO. Mutations in the tox operator sequence indicate that the binding of DtxR dimers to DNA may be highly cooperative.


We thank Dr. Ted J. Huston for the ICP–MS analyses and Stephen P. L. Cary for the SPR suggestion. We thank the Margolis lab (Michigan) and the Bertozzi lab (Berkeley) for the Biacore 3000 and the Marqusee lab (Berkeley) for the Aviv 62DS spectropolarimeter. The advice and suggestions of John F. Love (Boston University Medical School) are greatly appreciated. National Institutes of Health Grant CA26731 (to M.A.M.) supported this research. M.M.S. was supported by National Institutes of Health Grant T32 GM08597 and an American Foundation for Pharmaceutical Education Fellowship.


diphtheria toxin repressor
wild-type DtxR
3-(2-pyridyl)-5,6-bis(4-phenylsulfonic acid)-1,2,4-triazine
quantitative amino acid analysis
gel electrophoretic mobility-shift assay
inductively coupled plasma high-resolution mass spectrometer
surface plasmon resonance


This paper was submitted directly (Track II) to the PNAS office.


1. Pappenheimer A M, Jr, Johnson S J. Br J Exp Pathol. 1936;17:335–341.
2. Tai S P, Krafft A E, Nootheti P, Holmes R K. Microb Pathol. 1990;9:267–273. [PubMed]
3. Schmitt M P, Holmes R K. J Bacteriol. 1994;176:1141–1149. [PMC free article] [PubMed]
4. Lee J H, Wang T, Ault K, Liu J, Schmitt M P, Holmes R K. Infect Immun. 1997;65:4273–4280. [PMC free article] [PubMed]
5. Schmitt M P. Infect Immun. 1997;65:4634–4641. [PMC free article] [PubMed]
6. Schmitt M P. J Bacteriol. 1997;179:838–845. [PMC free article] [PubMed]
7. Schmitt M P, Talley B G, Holmes R K. Infect Immun. 1997;65:5364–5367. [PMC free article] [PubMed]
8. Boyd J, Oza M N, Murphy J R. Proc Natl Acad Sci USA. 1990;87:5968–5972. [PMC free article] [PubMed]
9. Schmitt M P, Holmes R K. Infect Immun. 1991;59:1899–1904. [PMC free article] [PubMed]
10. Schmitt M P, Holmes R K. Infect Immun. 1991;59:3903–3908. [PMC free article] [PubMed]
11. Boyd J M, Hall K C, Murphy J R. J Bacteriol. 1992;174:1268–1272. [PMC free article] [PubMed]
12. Schmitt M P, Twiddy E M, Holmes R K. Proc Natl Acad Sci USA. 1992;89:7576–7580. [PMC free article] [PubMed]
13. Tao X, Boyd J, Murphy J R. Proc Natl Acad Sci USA. 1992;89:5897–5901. [PMC free article] [PubMed]
14. Tao X, Murphy J R. J Biol Chem. 1992;267:21761–21764. [PubMed]
15. Schmitt M P, Holmes R K. Mol Microbiol. 1993;9:173–181. [PubMed]
16. Tao X, Murphy J R. Proc Natl Acad Sci USA. 1994;91:9646–9650. [PMC free article] [PubMed]
17. Qiu X, Verlinde C L, Zhang S, Schmitt M P, Holmes R K, Hol W G. Structure. 1995;3:87–100. [PubMed]
18. Schiering N, Tao X, Zeng H, Murphy J R, Petsko G A, Ringe D. Proc Natl Acad Sci USA. 1995;92:9843–9850. [PMC free article] [PubMed]
19. Ding X, Zeng H, Schiering N, Ringe D, Murphy J R. Nat Struct Biol. 1996;3:382–387. [PubMed]
20. Qiu X, Pohl E, Holmes R K, Hol W G. Biochemistry. 1996;35:12292–12302. [PubMed]
21. Pohl E, Qui X, Must L M, Holmes R K, Hol W G. Protein Sci. 1997;6:1114–1118. [PMC free article] [PubMed]
22. Pohl E, Holmes R K, Hol W G. J Biol Chem. 1998;273:22420–22427. [PubMed]
23. Wang G, Wylie G P, Twigg P D, Caspar D L, Murphy J R, Logan T M. Proc Natl Acad Sci USA. 1999;96:6119–6124. [PMC free article] [PubMed]
24. Twigg P D, Parthasarathy G, Guerrero L, Logan T M, Caspar D L. Proc Natl Acad Sci USA. 2001;98:11259–11264. [PMC free article] [PubMed]
25. White A, Ding X, vanderSpek J C, Murphy J R, Ringe D. Nature. 1998;394:502–506. [PubMed]
26. Pohl E, Holmes R K, Hol W G. J Mol Biol. 1999;292:653–667. [PubMed]
27. Bradford M M. Anal Biochem. 1976;72:248–254. [PubMed]
28. Fish W W. Methods Enzymol. 1988;158:357–364. [PubMed]
29. Stookey L L. Anal Chem. 1970;42:779–781.
30. Myszka D G, Morton T A. Trends Biochem Sci. 1998;23:149–150. [PubMed]
31. Crankshaw M W, Grant G A. In: Current Protocols in Protein Science. Chanda V B, Leonard S A, editors. Vol. 2. New York: Wiley; 2001. p. 15.1.13. (Suppl. 3).
32. Wang Z, Schmitt M P, Holmes R K. Infect Immun. 1994;62:1600–1608. [PMC free article] [PubMed]
33. Tao X, Zeng H Y, Murphy J R. Proc Natl Acad Sci USA. 1995;92:6803–6807. [PMC free article] [PubMed]
34. Goranson-Siekierke J, Pohl E, Hol W G, Holmes R K. Infect Immun. 1999;67:1806–1811. [PMC free article] [PubMed]
35. Chen C S, White A, Love J, Murphy J R, Ringe D. Biochemistry. 2000;39:10397–10407. [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...