Elucidation of the Functional Metal Binding Profile of a Cd^II/Pb^II sensor CmtR^Sc from Streptomyces coelicolor

Abstract

Metal homeostasis and resistance in bacteria is maintained by a panel of metal sensing transcriptional regulators that collectively control transition metal availability and mediate resistance to heavy metal xenobiotics, including As^III, Cd^II, Pb^II and Hg^II. The ArsR family constitutes a superfamily of metal sensors that appear to conform to the same winged helical, homodimeric fold, that collectively “sense” a wide array of beneficial metal ions and heavy metal pollutants. The genomes of many actinomycetes, including the soil dwelling bacterium Streptomyces coelicolor and the human pathogen Mycobacterium tuberculosis, encode over ten ArsR family regulators, most of unknown function. Here, we present the characterization of a homolog of M. tuberculosis CmtR (CmtR^Mtb) from S. coelicolor, denoted CmtR^Sc. We show that CmtR^Sc, in contrast to CmtR^Mtb binds two monomer mol equivalents of Pb^II or Cd^II to form two pairs of trigonal S₃ coordination complexes per dimer. Metal site 1 conforms exactly to the α4C site previously characterized in CmtR^Mtb while metal site 2 is coordinated by a C-terminal vicinal thiolate pair, Cys110 and Cys111. Biological assays reveal that only Cd^II and, to a lesser extent, Pb^II mediate transcriptional derepression in the heterologous host M. smegmatis in a way that requires metal site 1. In contrast, mutagenesis of metal site 2 ligands Cys110 or Cys111 significantly reduces Cd^II responsiveness, with no detectable effect on Pb^II sensing. The implications of these findings on the ability to predict metal specificity and function from metal-site “signatures” in the primary structure of ArsR family proteins are discussed.

The concentrations of first-row d-block transition metal ions and other heavier, toxic di- and trivalent metal ions and metalloids that play no biological role can vary dramatically in the immediate environment of a bacterial community. Although some are essential, and enable proteins to adopt their native three-dimensional structures or function as cofactors for metalloenzyme catalysis, an excess beyond that required by normal cellular metabolism can also be strongly deleterious to cell viability (1). As a result, metal-specific regulatory systems have evolved that detect metal-sufficiency or toxicity and control the intracellular availability of the different metal ions as well as purge any toxic elements or compounds from the cytosol. In free-living microorganisms, such detoxification/resistance systems are particularly important because it permits these organisms to survive by adaptation to potentially harsh environmental conditions, e.g., soil and water heavily contaminated by heavy metal salts (2) and/or the phagosome of mammalian host cells for certain microbial pathogens (3). Large, highly polarizable and thiophilic metal ions such as Hg^II, Cd^II and Pb^II are highly toxic, in part because they are capable of forming high affinity, kinetically long-lived complexes with protein cysteine thiolates found in native zinc binding sites, thereby rendering them nonfunctional. Such proteins include canonical zinc finger proteins and 5-aminolevulinic acid dehydratase (ALAD) in mammals (4, 5).

CmtR is a Cd^II sensing SmtB/ArsR (or ArsR) family metalloregulatory repressor from Mycobacterium tuberculosis (6–9). In addition to Cd^II, M. tuberculosis CmtR (denoted CmtR^Mtb here) senses Pb^II in the heterologous host M. smegmatis and has been shown to form Cys-thiolate rich metal coordination complexes with Cd^II, Pb^II and even Zn^II in vitro (6, 7). CmtR possesses a unique pair of α4C metal binding sites (metal ligands derived from the α4 helix and the C-terminal tail; see Fig. 1B) (6, 8, 10, 11) structurally distinct from the Cd^II/Pb^II sites of another SmtB/ArsR Cd^II/Pb^II sensor, Staphylococcus aureus plasmid pI258-encoded CadC, which contains functional α3N and nonfunctional α5 metal sites (1, 12, 13). The solution structure of the CmtR-Cd^II complex reveals a homodimer with metal bound to Cys102' from the C-terminal tail region of one subunit and Cys57, Cys61 in the helix α4 from the other subunit; Pb^II is predicted to bind to the same pair of sites (7, 8) (Fig. 1). Cys102 plays an accessory role in stabilizing the coordination complex while Cys57 and Cys61 anchor it, contributing most of the metal binding affinity (7). However, Cys102 does function as a key allosteric metal ligand in mediating the disassembly of oligomeric CmtR-cmt O/P¹ oligomeric complexes in vitro (7, 8), and for derepression in vivo (6), although the structural and dynamic changes induced in the dimer upon Cys-102 Cd^II binding remain to be characterized.

Figure 1. Genomic region of CmtR homologs in S. coelicolor, solution structure and multiple sequence alignment of M. tuberculosis CmtR.

(A) Genomic region around S. coelicolor CmtR^Sc (SCO0875) and SCO3522; each gene is separated by a single TGA termination codon (*) from homologous downstream genes SCO0874 and SCO3521 that encode putative CDF-family heavy metal transporters. The surrounding genomic regions are completely unrelated in the two loci. (B) Ribbon diagram of the solution structure of the M. tuberculosis CmtR-Cd^II complex (2jsc) with individual protomers shaded slate and violet and the two symmetry-related Cd^II ions colored green (8). The secondary structural units are labeled, with the side-chains of C24 and the Cd^II-coordinating residues, C57, C61 and C102' (prime designation, opposite subunit) highlighted in stick. The most C-terminal residue in the structural model of each protomer is R106. (C) Multiple sequence alignment of CmtR^Mtb, CmtR^Sc and the product of SCO3522. Cd^II-coordinating residues in CmtR^Mtb are highlighted in bold, with other residues as in panel B.

Recent in vivo experiments using M. tuberculosis suggest that CmtR^Mtb binds cooperatively to four binding sites in an extended 90-bp region upstream of the cmtR-Rv1993c-cmtA operon, to inhibit the interaction of RNA polymerase with the promoter region (9). These studies also suggest that Cd^II is the sole inducer in M. tuberculosis, a situation that contrasts with findings in the heterologous host, M. smegmatis (6, 9). The cmtA gene encodes a deduced metal transporting P_1B-ATPase efflux pump, which is proposed to efflux toxic metal ions from the bacterial cytosol (6, 9). Therefore, increasing cellular Cd^II concentration triggers derepression of the cmtR-Rv1993c-cmtA operon, resulting in increased concentrations of CmtA in the plasma membrane which is thought to export Cd^II from the cytosol against a metal concentration gradient (9).

Streptomyces coelicolor A3(2) is representative of a ubiquitous group of soil-dwelling, filamentous Gram-positive bacteria. This highly adaptable organism undergoes a broad range of metabolic processes and biotransformations, and is noted for its natural antibiotic production (14). Like M. tuberculosis, S. coelicolor belongs to the taxonomy order of Actinomycetales (14, 15). Although each has very different lifestyles, both encode elaborate metal detoxification and efflux systems. As a human pathogen, M. tuberculosis must adapt to various microenvironmental niches, including the phagosome of infected macrophages where it encounters changes in metal availability (3, 16–19). S. coelicolor must also sense and respond to fluctuating metal levels within soils (20–22). Notably, the sequenced strains of M. tuberculosis and S. coelicolor possess multiple deduced ArsR family regulators, ten (10) and fourteen, respectively, as well as representatives of many of the other metalloregulatory protein classes (1, 23). The diversity of the metal-responsive regulators in these two organisms is therefore likely to reflect their different ecological niches and the different survival strategies employed to avoid metal stress. A search for protein homologs of M. tuberculosis CmtR returned two open reading frames in S. coelicolor A3(2) corresponding to locus tags SCO0875 and SCO3522. The gene products of SCO0875 and SCO3522 possess only ten amino acid differences between them, and each shares ≈50% identity with M. tuberculosis CmtR. Both SCO0875 and SCO3522 are immediately upstream of deduced cation diffusion facilitator (CDF) family integral membrane metal transporters (Fig. 1A), with predicted roles in the transport of the more thiophilic metal ions (24). SCO0875 was chosen for detailed study herein and designated CmtR^Sc.

CmtR^Sc shares four conserved cysteine residues with CmtR^Mtb which include the three Cd^II ligands (Cys57, Cys61 and Cys102) required for metal binding and allosteric regulation (6) as well as C24 in the α2 helix (Fig. 1A). As a result, we expected to observe at least partially similar metal binding profiles for CmtR^Mtb and CmtR^Sc. However, CmtR^Sc contains two additional cysteines arranged as C-terminal vicinal pair (Cys110 and Cys111) that could also be involved in metal coordination (Fig. 1C). It was therefore of interest to functionally and structurally characterize CmtR^Sc and compare its properties to that of CmtR^Mtb. We show here that the CmtR^Sc homodimer harbors a second pair of high affinity Cys-thiolate rich Cd^II/Pb^II coordination sites relative to CmtR^Mtb that involves metal coordination by Cys110 and Cys111. We further show that wild-type CmtR^Sc, is a bona fide CmtR that is highly selective for Cd^II/Pb^II in the heterologous host M. smegmatis, the same host used to characterize CmtR^Mtb, with no other metal ions, including Hg^II and Zn^II, capable of inducing transcriptional derepression. Strikingly, the second pair of Cd^II/Pb^II sites unique to CmtR^Sc is required for full Cd^II-responsiveness in M. smegmatis, but does not alter Pb^II sensing under the same conditions. The structural and functional implications of these findings for the prediction of metal-sensing sites in ArsR family metal sensors are discussed.

MATERIALS AND METHODS

Construction of wild-type and C110G/C111S CmtR^Sc overexpression plasmids

To create pET3a-CmtR^Sc, the CmtR^Sc coding region was amplified by PCR from Streptomyces violaceoruber genomic DNA (M145, used in to sequence the S. coelicolor genome (14) using primers V (5'-AAAAACATATGGTGCTGACTCTCGCTGCCGATATC-3') and VI (5'-AAAAAGCTAAGCTCAGCAGCACTCCTTCTCGTC-3') and cloned into pET3a (Novagen) between the NdeI and BPU1102I restriction sites using standard methods. The plasmid encoding C110G/C111S CmtR^Sc was generated by site-directed mutagenesis using the QuikChange kit (Stratagene) and primers VII (5'-GGACGAGAAGGAGGGCTCCTGAGCAATAACTAGC-3') and primer VIII (5'-GCTAGTTATTGCTCAGGAGCCCTCCTTCTCGTCC-3') using pET3a-CmtR^Sc as the template. The integrity of all expression plasmids was verified by DNA sequencing.

Purification of wild type and mutant CmtR^Sc

CmtR^Sc and C110G/C111S CmtR^Sc were purified to homogeneity using a procedure previously described for CmtR^Mtb (7). The theoretical molecular weight for CmtR^Sc is 12.2 kDa and the extinction coefficient was calculated to be 3730 M⁻¹•cm⁻¹.

Free thiol determination

A standard DTNB¹ colorimetric assay was used to determine the number of free thiols in CmtR^Sc (25). 25 μL 2.5 mM DTNB solution was added separately into 400 μL 10–15 μM protein. After a 30 min incubation in the anaerobic chamber, the concentration of thiolate anion was quantified at 412 nm (ε=13,600 M⁻¹•cm⁻¹) after subtraction of the absorbance of the final dialysis buffer with same concentration of DTNB. The number of free thiols in wild-type CmtR^Sc was 5.8 ± 0.2 (6 expected) and 3.0 ± 0.2 (4 expected) for C110G/C111S CmtR^Sc.

Atomic absorption spectroscopy

The concentrations of all metal titrants were determined using a Perkin-Elmer Analyst 700 atomic absorption spectrophotometer operating in flame mode using a different hollow cathode lamps specific for each metal (25). Zn^II was detected at 213.9 nm (slit=0.7 nm), Cd^II was detected at 228.8 nm (slit = 0.7 nm) and Pb^II was detected at 283.3 nm (slit = 0.7 nm).

Analytical Sedimentation Equilibrium Ultracentrifugation

All experiments were carried out on a Beckman Optima XL-A analytical ultracentrifuge with the rotor speed set to 20,000 rpm at 25.0 °C. Ultracentrifuge cells were assembled in the anaerobic glove box and contained 5.0 μM CmtR^Sc wild-type or 8.0 μM C110G/C111S CmtR^Sc in buffer (10 mM Hepes, 0.4 M NaCl, and 0.1 mM EDTA at pH 7.0). Scans were monitored by absorbance at 232 nm or 233 nm for the wild-type and C110G/C111S CmtR^Sc, respectively, with the final seven scans extracted and subjected to simultaneous fitting using Ultrascan II 8.0. A partial specific volume (ν) of 0.7354 ml/g (predicted by Sednterp 1.07 software, www.bbri.org/RASMB/rasmb.html) and a buffer density (ρ) of 1.0 g/ml were used in the analysis. The data were globally and simultaneously fitted to either a one component ideal species model or a monomer-dimer equilibrium model using eqs. 1 and 2, respectively:

$$C\left(\mathrm{X}\right)=e^{\left[\frac{\ln \left(A\right)+M{\omega }^2(1-\stackrel{‒}{\nu }\rho )(X^2-X_r^2)}{2RT}\right]}+B$$ (Eq. 1)

$$C\left(\mathrm{X}\right)=e^{\left[\frac{2\ln \left(A\right)+\ln \left(\frac{2}{EL}\right)+\ln \left(K_{1,2}\right)+2M{\omega }^2(1-\stackrel{‒}{\nu }\rho )(X^2-X_r^2)}{2RT}\right]}+e^{\left[\frac{\ln \left(A\right)+M{\omega }^2(1-\stackrel{‒}{\nu }\rho )(X^2-X_r^2)}{2RT}\right]}+B$$ (Eq. 2)

where X = cell radius, X_r = reference radius, A = amplitude of monomer*, M = molecular weight of monomer*, E = extinction coefficient, R = gas constant, T = temperature, B = baseline offset*, ω = angular velocity, L = optical pathlength, K_1,2 = monomer–dimer equilibrium constant*, and * indicates this parameter can be floated during parameter optimization (26). Fits to the monomer-dimer model did not significantly improve the goodness-of-fit, consistent with a low concentration of monomeric species under these solution conditions; thus, K_1,2≥10⁶ M⁻¹.

Cd^II, Pb^II optical absorption spectroscopy

All metal binding experiments were carried out anaerobically at ambient temperature (~25°C) using a Hewlett-Packard model 8452A spectrophotometer (12, 27). For Cd^II titrations, apo proteins were diluted using final dialysis buffer S to ≈ 50 μM in 800 μL and loaded into an anaerobic cuvette fitted with a Hamilton gas-tight adjustable volume syringe containing 0.5 mM Cd^II titrant in the glove box. Complete optical spectra of apo protein were collected from 200–900 nm 1–2 min after each addition of a known aliquot (5–15 μL) of Cd^II. Corrected spectra were obtained by subtraction of the apo protein spectra from each spectrum obtained after addition of metal ion, and further corrected for dilution. Pb^II titrations were done in exactly the same way except that 10 mM bis-Tris, 0.4 M NaCl, pH 7.0 was used as the buffer. Bis-Tris is a weakly chelating buffer that can prevent excess Pb^II from forming Pb^II(OH)₂ precipitate (4). Cd^II competition experiment performed with EDTA was carried out as described above except 200–300 μM chelator was present. The competition binding curves were fit to the model using Dynafit:

$$\mathrm{P}+\mathrm{Cd}\leftrightarrows \mathrm{P}.\mathrm{Cd},K_{\mathrm{Cd}1};$$

$$\mathrm{P}.\mathrm{Cd}+\mathrm{Cd}\leftrightarrows \mathrm{P}.\mathrm{Cd}2,K_{\mathrm{Cd}2};$$

$$\mathrm{EDTA}+\mathrm{Cd}\leftrightarrows \mathrm{EDTA}.\mathrm{Cd},K_{\mathrm{Cd}\text{-}\mathrm{EDTA}}$$

where P is the monomer concentration of CmtR^Sc and K_Cd-EDTA (pH 7.0) = 8.31 × 10¹³ M⁻¹ under these conditions using a pH-dependent log K_Cd=16.5 and pK_as of 9.52 and 6.13 for EDTA (28).

Zn^II binding experiments

The zinc chelator dye magfura-2 (K_Zn = 5.0 × 10⁷ M⁻¹ at pH 7.0 and 25°C) was used as a colorimetric competitor for Zn^II binding by wild-type and C110G/C111S CmtR^Sc as previously described (29). 30 μM CmtR^Sc and ≈15 μM magfura-2 was used for all experiments. The data were fit to a competitive binding model assuming two nonequivalent, noninteracting metal binding sites per CmtR^Sc protomer for CmtR^Sc wild-type by DynaFit:

$$\mathrm{P}+\mathrm{Zn}\leftrightarrows \mathrm{P}-\mathrm{Zn},K_{\mathrm{PZn}};$$

$$\mathrm{P}.\mathrm{Zn}+\mathrm{Zn}\leftrightarrows \mathrm{P}-{\mathrm{Zn}}_2,K_{\mathrm{PZn}2};$$

$$\mathrm{magfura}\_2+\mathrm{Zn}\leftrightarrows \mathrm{magfura}\_2-\mathrm{Zn},$$

where K_Zn-magfura2 = 5.0 × 10⁷ M⁻¹ under these conditions. For C110G/C111S CmtR^Sc, data were fitted as above to a competitive binding model assuming one metal binding site per protomer or two per dimer (without K_PZn2).

Construction of promoter-lacZ fusions, site-directed mutagenesis and β-galactosidase assays

CmtR^Sc upstream sequences and the coding region (539 bp) were amplified from S. coelicolor genomic DNA using primers 5'- GAAAGTACTGCGGGTAGTGGCGATGTGATCC-3' and 5'-GAAGGTACCGCGGTCATTCAGCAGC-3', and ligated to pGEM-T prior to subcloning into the ScaI/KpnI site of pJEM15 (30) to create a transcriptional fusion with lacZ. QuickChange XL (Stratagene) site-directed mutagenesis was subsequently employed to generate derivatives with the following codon substitutions in CmtR^Sc: Cys24, Cys57, Cys61, Cys102, Cys110, and Cys111, each to Ser; and Arg16 to a UGA stop codon. All generated plasmid constructs were checked by sequence analysis. Mycobacterium smegmatis mc²155 was used as an actinomycete host for reporter gene assays. The lacZ fusion constructs were introduced into M. smegmatis and transformants selected on Luria-Bertani (LB) agar supplemented with 25 μg mL⁻¹ kanamycin as described previously (30). β-galactosidase assays were performed as described (31) in triplicate on at least three separate occasions. Cells were grown at 37 °C with shaking in LB broth or Sauton medium containing 0.05% (v/v) Tween 80 and kanamycin (25 μg ml⁻¹) supplemented with the indicated concentration of metal salt (described in individual experiments) for ~20 h immediately prior to assays. The metal salts used were ZnSO₄, CoSO₄, NiCl₂, CdCl₂, Pb(C₂H₃O₂)₂, CuSO₄, AsNaO₂ and HgCl₂.

RESULTS

Wild-type and C110G/C111S CmtR^Sc are stable homodimers at low protein concentrations

Both CmtR^Sc and its C110G/C111S mutant were subjected to analytical equilibrium sedimentation ultracentrifugation at 20,000 rpm at 5 μM and 8 μM monomer concentration, respectively, and absorbance recorded at 232 nm, with representative results shown in Fig. 2. Each set of experimental data that had reached equilibrium was first subjected to a global simultaneous fit to a single ideal species model indicated by a solid line. The apparent molecular weights obtained from these data (Fig. 2) are consistent with both apo-CmtR^Sc and apo-C110G/C111S CmtR^Sc being predominantly homodimeric under these conditions.

Figure 2. Analytical sedimentation equilibrium ultracentrifugation of CmtR^Sc and C110G/C111S CmtR^Sc.

(A) 5.03 μM monomer wild-type CmtR^Sc. (B) 8.0 μM monomer C110G/C111S CmtR^Sc. Filled symbols in upper panels represent an overlay of data collected during the last seven scans and indicate that equilibrium had been reached. The solid line represents the global simultaneous fit for a single ideal species model using Ultrascan. For wild-type CmtR^Sc, the fitted M_w is 23,820 Da (theoretical dimer M_w = 24,378 Da), variance = 1.0689 e⁻⁴. For C110G/C111S CmtR^Sc, the fitted M_w is 25,690 Da (theoretical dimer M_w = 24,252 Da), variance = 5.8026 e⁻⁵. Conditions: 10 mM Hepes, 0.4 M NaCl, and 0.1 mM EDTA at pH 7.0, 25.0 °C, and 20,000 rpm rotor speed.

CmtR^Sc possesses two Cd^II binding sites per monomer while C110G/C111S CmtR^Sc possesses one

Apoprotein-subtracted Cd^II absorbance spectra of CmtR^Sc are shown in Fig. 3. The saturated Cd^II spectra is characterized by an intense ligand-to-metal charge transfer (LMCT¹) at 240 nm, similar to that observed for CmtR^Mtb (7). Binding isotherms (insets) were obtained by plotting the corrected absorbance at 240 nm as a function of total [Cd^II] over protein monomer ratio. The maximum molar absorptivity at 240 nm for CmtR^Sc is ≈ 31,000 M (monomer)⁻¹ cm⁻¹ which is two-fold larger than wild-type mycobacterial CmtR^Mtb (ε ≈ 16,000 M⁻¹ cm⁻¹) (Fig. 3A) (7). In addition, CmtR^Sc exhibits a 2:1 (Cd^II:CmtR^Sc monomer) binding stoichiometry (Fig. 3A), in contrast to CmtR^Mtb which is known to bind one metal per protomer (7). Nonlinear least-squares fits of the binding isotherms to a simple 2:1 independent-site metal binding model returns only a lower limit of the binding affinity given the stoichiometric nature of these binding curves (K_Cd > ≈5 × 10⁷ M⁻¹) (Fig. 3A).

Figure 3. Cd^II titrations of wild-type and C110G/C111S CmtR^Sc.

(A) Apoprotein subtracted difference spectrum of wild-type CmtR^Sc (50.7 μM monomer). (B) Apoprotein subtracted difference spectrum of C110G/C111S CmtR^Sc (44.8 μM monomer). CmtR^Sc variants were titrated anaerobically with increasing concentrations of Cd^II. Inset, Cd^II binding isotherm plotted as change in A₂₄₀ vs. [CmtR^Sc variant monomer]. (C) Cd^II-EDTA competition binding isotherm in which 20.9 μM wild-type CmtR^Sc was titrated with Cd^II in the presence of 227 μM EDTA. Conditions: 10 mM Hepes, 0.4 M NaCl, pH 7.0, 25°C.

The presence of an additional Cd^II site in the CmtR^Sc homodimer coupled with twice the Cys S⁻→Cd^II molar absorptivity prompted us to investigate if the two vicinal cysteine residues (Cys110 and Cys111) at the C-terminus donate thiolate ligands to the second Cd^II ion. We measured the Cd^II absorption spectrum of C110G/C111S CmtR^Sc, where the two C-terminal cysteines were converted to their analogous residues in CmtR^Mtb (Fig. 1C). As expected, C110G/C111S CmtR^Sc behaves much like CmtR^Mtb, given a ≈0.9:1 Cd^II:C110G/C111S monomer stoichiometry and an identical monomer molar absorptivity of ≈ 16,000 M⁻¹ cm⁻¹ (Fig. 3B) (7), which is precisely half that of wild-type CmtR^Sc. As expected, the binding of Cd^II is stoichiometric under these conditions and thus returns only a lower limit of the metal binding affinity (Fig. 3B). Given a molar absorptivity of ≈5500 M⁻¹ cm⁻¹ per Cd-S coordination bond (32–34), it is reasonable to conclude that C110G/C111S CmtR^Sc possesses one Cd^II site (termed metal site 1) that is identical to the previously characterized canonical α4C site CmtR^Mtb, while wild-type CmtR^Sc possesses two spectroscopically similar Cd^II binding sites (metal sites 1 and 2).

The Cd^II affinities of each metal site per monomer in wild-type CmtR^Sc were estimated by carrying out a Cd^II titration experiment in the presence of a known concentration of the metal chelator EDTA (Fig. 3C), with K_Cd-EDTA = 8.3 × 10¹³ M⁻¹ under these solution conditions (pH 7.0, 25.0 °C). A nonlinear least squares fit of a simple competition model assuming two non-equivalent, non-interacting sites per protomer gives K_Cd1 = 3.0 (±0.1) × 10¹³ M⁻¹, K_Cd2 = 3.6 (±0.5) × 10¹² M⁻¹. These data taken together confirm that CmtR^Sc has two pairs of structurally similar high affinity metal sites per homodimer, which differ in macroscopic affinity by ≈10-fold. Previous findings with CmtR^Mtb reveal that Cys57, Cys61 and Cys102 are ligands to Cd^II and form trigonal pyramidal coordination geometry (S₃) or distorted tetrahedral coordination geometries (S₃O), with Cys102 not as strongly bound as the other thiolate ligands (6–8) (see Fig. 1B). Considering these three cysteines are conserved in CmtR^Mtb and CmtR^Sc, they likely create a Cd^II site indistinguishable to that in CmtR^Mtb and is denoted metal site 1. In contrast, Cd^II site 2 clearly requires coordination by Cys110 and Cys111.

CmtR^Sc possesses two Pb^II binding sites and the same C110G/C111S mutation abolishes one

Anaerobic Pb^II titration experiments were performed using ≈50 μM apo-wild-type CmtR^Sc (Fig. 4A) and C110G/C111S CmtR^Sc (Fig. 4B) and the apoprotein-subtracted difference spectra are shown. The saturated Pb^II spectra of both proteins are characterized by an intense absorption in the far-ultraviolet and a long-wavelength absorption band with maximum at 333 nm, identical to that of CmtR^Mtb (7) and report on ligand-to-metal charge transfer (S⁻ 3p→Pb 6p) and intraatomic (Pb 6s²→Pb 6sp) electronic transitions (35, 36). Binding isotherms (insets) were obtained by plotting the corrected absorbance at 333 nm as a function of total [Pb^II]/protein monomer ratio. As expected, for wild-type CmtR^Sc Pb^II binding is saturable at ≈ 2:1 ratio over [CmtR^Sc] monomer (Fig. 4A); in contrast, this stoichiometry drops to ≈ 0.7 Pb^II/monomer for C110G/C111S CmtR^Sc (Fig. 4B). These Pb^II stoichiometries are largely consistent with the relative monomer molar absorptivities at 333 nm (ε₃₃₃) of ≈7900 M⁻¹ cm⁻¹ and ≈3950 M⁻¹ cm⁻¹ for wild-type CmtR^Sc and C110G/C111S CmtR^Sc, respectively (Fig. 4). Interestingly, the Pb^II-thiolate molar absorptivity of C110G/C111S CmtR^Sc appears to be about one-half that of CmtR^Mtb (7) for reasons that are not clear, although the fractional stoichiometries in each case may complicate this. Furthermore, unlike for Cd^II-binding, the extent to which molar absorption wavelength and intensity reports on coordination number and geometry for Pb^II-thiolate complexes has not yet been firmly established (12, 36–39). In any case, the simplest conclusion is that substitution of Cys110 and Cys111 with non-liganding residues abolishes Pb^II binding to metal site 2, as is the case for Cd^II.

Figure 4. Pb^II titrations of CmtR^Sc and C110G/C111S CmtR^Sc.

(A) Apoprotein subtracted difference spectrum of wild-type CmtR^Sc (50.8 μM). (B) Apoprotein subtracted difference spectrum of C110G/C111S CmtR^Sc (53.7 μM). CmtR^Sc variants were titrated anaerobically with increasing concentrations of Pb^II. Inset, Pb^II binding isotherm plotted as change in A₃₃₃ vs. [CmtR^Sc variant monomer] Conditions: 10 mM Bis-Tris, 0.4 M NaCl, pH 7.0, 25.0 °C.

CmtR^Sc binds two Zn^II per monomer as determined by chelator competition experiments with magfura-2

The zinc indicator dye magfura-2 (K_{Zn•mag-fura-2} = 5.0 × 10⁷ M⁻) was used as a competitor of Zn^II binding to CmtR^Sc and the C110G/C111S mutant to determine the Zn^II binding stoichiometry and affinity constant K_Zn (11). Figure 5 shows representative titrations of Zn^II into mixtures of wild-type CmtR^Sc (Fig. 5A) and C110G/C111S CmtR^Sc (Fig. 5B) and magfura-2. The solid curve represents a fit to a model that describes the binding of two Zn^II ions to the wild-type CmtR^Sc monomer, considering each monomer-metal binding as an independent event (Fig. 5A). For wild-type CmtR^Sc, the estimated parameters are similar for both zinc ions, with K_PZn = 5.3 (±1.8) ×10⁸ M⁻ and K_PZn2 = 6.7 (±1.4) ×10⁸ M⁻. These binding affinities differ from that determined previously for the CmtR^Mtb dimer, which showed strong negative cooperativity within the dimer with one site in ≈10¹⁰ M⁻ range and the other in ~10⁵ M⁻¹ range under the same solution conditions (7). For the C110G/C111S CmtR^Sc mutant, competition titration curves were fit to a model that describes the binding of one Zn^II ion to the C110G/C111S CmtR^Sc monomer (Fig. 5B), with an estimated affinity of K_PZn = 5.5 (±1.2) × 10⁷ M⁻¹.

Figure 5. Zn^II titrations of CmtR^Sc and C110G/C111S CmtR ^Sc using magfura-2 as an indicator.

(A) 30 μM CmtR^Sc and 14.7 μM magfura-2 and (B) 30 μM C110G/C111S CmtR^Sc and 15 μM magfura-2 were present. The empty circles represent A₃₂₅ and the filled squares represent A₃₆₆. The solid line represents a global non-linear least square fit to a model that incorporates the stepwise binding of two Zn^II (defined by K_PZn and K_PZn2) to a CmtR^Sc monomer using Dynafit. The titration for C110G/C111S CmtR^Sc was fitted using one Zn^II to protein monomer binding model. The following parameters were obtained for wild-type CmtR^Sc: K_PZn = 5.3 (±1.8) × 108 M⁻¹ (a lower limit under these conditions), K_PZn2 = 6.7 (±1.4) × 10⁸ M⁻¹. For C110G/C111S CmtR^Sc, K_PZn = 5.5 (±1.2) × 10⁷ M⁻¹. Conditions: 10 mM Hepes, 0.4 M NaCl at pH 7.0 and 25.0 °C.

Cadmium and lead alleviate CmtR^Sc mediated repression

M. smegmatis mc²155 was previously used to characterize CmtR^Mtb and hence was exploited as a heterologous host for the characterization of CmtR^Sc, thereby allowing direct comparison of these two regulators within the same cytosol. M. smegmatis mc²155 and S. coelicolor are closely related actinomycetes, with similarly high GC-contents, and recognition of Streptomyces operator-promoter elements by the mycobacterial transcription machinery (40, 41) has been exploited previously to characterize S. coelicolor genes in vivo (41). To determine which, if any, metals are sensed by CmtR^Sc in vivo, a 539-bp DNA fragment including the cmtR^Sc operator-promoter and coding region was fused to a promoterless lacZ in plasmid pJEM15 and introduced into M. smegmatis mc²155. β-galactosidase activity was measured following growth (~20 h) of cells in medium supplemented with maximum permissive concentrations of various metals. At these biologically significant metal levels, elevated activity was detected in response to Cd^II and, to a lesser extent, Pb^II but no other metals (Fig. 6A). Furthermore, in the absence of added metal ions, elevated β-galactosidase activity was detected in cells containing an analogous construct in which codon-16 within the cmtR^Sc coding region was converted to a stop codon (Fig. 6B), confirming that CmtR^Sc acts negatively toward expression.

Figure 6. CmtR^Sc responds to Cd^II and Pb^II in an actinomycete host.

(A) β-galactosidase activity measured in M. smegmatis mc²155 containing cmtR^Sc and its operator-promoter region fused to lacZ following growth in LB medium with no metal supplement or with maximum permissive concentrations of Zn^II (100 μM), Co^II (200 μM), Ni^II (500 μM), Cd^II (7.5 μM), Cu^II (500 μM), Pb^II (3.75 μM), As^III(20 μM), or Hg^II (0.025 μM). (B) β-galactosidase activity in cells containing wild-type CmtR^Sc (WT) or the stop codon derivative (R16*) following growth in LB medium with no metal supplement (black) or maximum permissive concentrations of Cd^II (gray) or Pb^II (white).

CmtR^Sc senses Cd^II and Pb^II using α4C sites

The previously characterized Cd^II binding α4C sensing sites of CmtR^Mtb involve Cys102 from the C-terminal region of one subunit in association with Cys57 and Cys61 from helix αR of the other subunit (6–8) (see Fig. 1B). Comparison of the amino acid sequences of CmtR^Sc and related ArsR family sensors reveals that the α4C ligands are completely conserved in CmtR^Sc and CmtR^Mtb, while CmtR^Sc lacks residues corresponding to other previously defined ArsR family metal-binding motifs (1, 10). Substitution of the α4C cysteines (Cys57, Cys61 and Cys102) in CmtR^Sc with serines created functional repressors that mediated low expression of lacZ in cells grown with no metal supplement, but repression was not alleviated at Cd^II concentrations that caused loss of repression by wild-type CmtR^Sc (Fig. 7A). In addition, no alleviation of repression was observed in the presence of maximum permissive concentrations of Pb^II (Fig. 7A). Hence, consistent with findings for CmtR^Mtb (6), Cys57, Cys61 and Cys102 at α4C are obligatory for both Cd^II and Pb^II recognition by CmtR^Sc in vivo and likely contribute toward a single common metal-binding site, metal site 1.

Figure 7. Metal sensing ligands of CmtR^Sc.

(A) β-galactosidase activity measured in M. smegmatis mc²155 containing wild-type CmtR^Sc (WT) or derivatives with indicated cysteine to serine codon substitutions, following growth in LB medium with no metal supplement (black) or maximum permissive concentrations of Cd^II (gray) or Pb^II (white). (B) β-galactosidase activity in cells containing wild-type CmtR^Sc (black circles) or the C110S (gray squares) and C111S derivatives (gray diamonds) grown in LB with up to inhibitory concentrations of Cd^II or Pb^II. Data points represent the mean (± SE) for three independent experiments, each performed in triplicate.

Cd^II, but not Pb^II, sensing by CmtR^Sc involves an additional pair of C-terminal cysteines

As discussed above, in addition to Cys57, Cys61 and Cys102 at α4C, CmtR^Sc possesses three further cysteines (Cys24 found in both CmtRs, and Cys110 and Cys111 which are unique to CmtR^Sc, Fig. 1C). In the absence of added Cd^II or Pb^II, β-galactosidase activity remained low in cells containing Ser substitutions of Cys24, Cys110 or Cys111, consistent with retention of repressor function (Fig. 7A). However, although inducer recognition was clearly retained in these cells, Cd^II-mediated derepression was substantially reduced in cells containing the single Cys110 and Cys111 substitution mutants (Fig. 7A). Indeed, when expression from the cmtR operator-promoter was examined in cells exposed to a range of Cd^II and Pb^II concentrations, Cd^II-responsiveness was significantly impaired for the C110S and C111S mutants at all cadmium concentrations up to inhibitory levels, whereas Pb^II-responsiveness appeared unaffected (Fig. 7B). Equivalent findings were also obtained using minimal (Sauton) medium (Fig. 8) which reveal no significant reduction in the magnitude of derepression by Pb^II in the C110S and C111S mutants, compared to wild-type CmtR^Sc, whereas Cd^II-responsiveness is substantially reduced. These results, when considered in the context of the metal binding properties of wild-type and C110G/C111S CmtR^Sc, reveal that coordination to metal site 2 in CmtR^Sc is required for full transcriptional derepression by Cd^II in the cell. In contrast, Pb^II is a less potent inducer of cmt-lacZ expression in vivo and does not require metal site 2 for this activity.

Figure 8. Metal sensing ligands of CmtR^Sc as determined on a chemically defined minimal medium.

Top panel, Cell viability of M. smegmatis mc²155 containing wild-type CmtR^Sc (black circles), C110S (gray squares) and C111S derivatives (gray diamonds) grown on minimal Sauton medium (containing 2.9 mM phosphate) as a function of total added Cd^II (left) or Pb^II (right). Lower panel, β-galactosidase activity in cells containing wild-type CmtR^Sc (black circles), C110S (gray squares) or C111S CmtR^Sc (gray diamonds) grown in minimal media up to inhibitory concentrations of Cd^II or Pb^II. Data points represent the mean (± SE) for three independent experiments, each performed in triplicate.

DISCUSSION

CmtR^Sc (SCO0875) is a M. tuberculosis CmtR homolog in S. coelicolor that shares 52% sequence identity with CmtR^Mtb. S. coelicolor belongs to the same taxonomic order (Actinomycetales) as the causative agents of tuberculosis and leprosy (M. tuberculosis and M. leprae), but is a soil-dwelling bacterium that may more often encounter heavy metal polluted environments (14). However, M. tuberculosis is also suggested to encounter heavy metal ions such as Cd^II due to their accumulation in alveolar macrophages as a result of air pollution and cigarette smoking (9). As a result, metal-specific regulatory machinery is required to manage the intracellular concentrations of these ions (1). M. tuberculosis CmtR is the founding member of a subfamily of winged helical DNA-binding repressors from the ArsR metal-sensing family that employ characteristic α4C metal sites that respond selectively to Cd^II and Pb^II (6–8). A search of CmtR^Mtb homologs in S. coelicolor identified two nearly identical open reading frames (locus tags SCO0875 and SCO3522) that appear to encode a CmtR^Sc that differ by just ten amino acids, and immediately upstream of ORFs that encode what are predicted to be nearly identical cation diffusion facilitator (CDF) family heavy metal transporters. The presence of what would appear to be two functionally redundant operons may well reflect the need for S. coelicolor to rapidly sense and efflux (or otherwise detoxify) any Cd^II and Pb^II encountered in the environment.

In this study, we chose the gene encoded by locus tag SCO0875 as representative of CmtR^Sc for detailed study and comparison with CmtR^Mtb. We were particularly intrigued by the presence of two C-terminal vicinal cysteines in CmtR^Sc, Cys110 and Cys111, both of which are conserved in SCO0875 and SCO3522. Substitution of these residues with non-liganding residues was expected to yield a regulator that behaves much like wild-type CmtR^Mtb. In vitro metal titration experiments confirm this, but also reveal a second metal site in CmtR^Sc which is formed by Cys-S⁻ coordination bonds donated by Cys110, Cys111 and a third as yet unidentified ligand. Cys24, which is at least partly exposed to solvent in the solution structure of Cd^II-bound CmtR (Fig. 1B) (8) could readily complete an S₃ or S₃(O) coordination complex. Since the solution structural model of Cd^II-bound CmtR ends at Arg106, this suggests that the C-terminal tail is highly flexible; if so, inspection of this structure suggests that the C-terminal Cys110/Cys111 pair of one subunit could potentially come in close physical proximity of Cys24 in the α2 helix of the same subunit by crossing over the “front” of the molecule in the orientation shown in Fig. 1B. In this case however Cys24 can not function as a key regulatory residue since in vivo β-galactosidase assays reveal that substitution of Cys24 does not influence Cd^II responsiveness of CmtR^Sc in contrast to Cys110 and Cys111; a similar finding characterizes the primary α4C metal site in C24S CmtR^Mtb (6). Cys24 may still complete the coordination structure of metal site 2, but it may simply increase the affinity of the site for Cd^II rather than function as an allosteric residue (23), much like that found previously for His100 in the Zn^II sensor S. aureus CzrA (42) and the Cys7 and Cys58 of the S₄ Cd^II sensor S. aureus pI258 CadC (12). Other non-thiolate possibilities for the third ligand are of course formally possible, but would not be compatible with the absorption spectroscopy of the Cd^II and Pb^II complexes (12). Detailed structural or ¹¹³Cd NMR studies (7, 12, 32) will be required to further elucidate the coordination structure of metal site 2 in CmtR^Sc.

CmtR^Sc displays an in vivo metal specificity for Cd^II and Pb^II that is identical to that of CmtR^Mtb in the heterologous host M. smegmatis; given the similarity of these two bacterial species, these findings suggest that CmtR^Sc is a bona fide Cd^II/Pb^II sensor in S. coelicolor. Although both CmtR^Sc and CmtR^Mtb share the same primary allosteric α4C metal site with the Hg^II sensor MerR in Streptomyces lividans (43), neither CmtR senses Hg^II in M. smegmatis (7). In addition, neither CmtR^Sc- nor CmtR^Mtb-repressed transcription is inducible by Zn^II (7), despite the fact that each retains the ability to bind Zn^II in vitro (7). Little is known about how Streptomyces spp. handle Hg^II toxicity and how this might differ from Mycobacteria; as a result, it is formally possible that in addition to detecting Cd^II and Pb^II, CmtR^Sc may detect Hg^II in the native host. A more important issue is how CmtRs sense Cd^II and Pb^II over Zn^II in the cell. Although absolute metal affinity of the sensing sites may not fully explain metal selectivity in the cell, this simple explanation may be at least partially operative here. Both proteins have Zn^II affinity ≤ 10¹⁰ M⁻¹, which is three orders less than the affinity for Cd^II (Fig. 3) (7) which itself is comparable to that previously measured K_Cd for the Cd^II/Pb^II sensor CadC (27). Furthermore, a K_Zn ≤10¹⁰ M⁻¹ may not be sufficient to be detected in cells by CmtR, since the typical bacterial Zn^II sensor is characterized by an equilibrium affinity for Zn^II in the 10¹²–10¹⁵ M⁻¹ range. In fact, a protein encoded by locus tag SCO6459 is a strong candidate α5 site zinc sensor in S. coelicolor that may be poised to detect weakly chelated Zn^II buffered in the 10⁻¹² to 10⁻¹⁵ M range (29, 42, 44, 45). This ensures that both cmtR^Sc-sco0874 and sco3522-sco3521 operons will be induced by Cd^II and Pb^II and not by Zn^II so that these toxic metals can be selectively effluxed from the cytosol.

The primary α4C metal site 1 is necessary for both Cd^II and Pb^II detection in the cell, but is not sufficient for full Cd^II responsiveness, which clearly requires coordination by metal site 2 (Figs. 7–8). The evolutionary advantage of this to S. coelicolor is unknown, but may reflect the possibility that Cd^II salts and low molecular weight complexes are far more bioavailable (soluble) in soils than Pb^II complexes (which may be readily precipitated) and thus must be more efficiently detoxified. Clearly, Cd^II is a stronger inducer of CmtR^Sc-mediated transcriptional derepression in M. smegmatis than is Pb^II over a similar concentration range of added metal, and most of this difference can be traced to Cd^II occupancy of metal site 2 (Figs. 6–7). Pb^II, being the larger cation, may well fail to induce the same change in homodimer structure and/or dynamics that has been shown to be important in CmtR^Mtb (8) and other ArsR family sensors (46). Alternatively, the bioavailable concentration of Pb^II achievable in M. smegmatis at the maximum permissive concentration of this cation (≈5 mM; Figs. 7–8) may not high enough to fill metal site 2 in the M. smegmatis cytosol; as a result, CmtR^Sc is a poorer Pb^II sensor in vivo. In any case, filling both metal sites in CmtR^Sc may well stabilize the allosterically inhibited low-affinity DNA binding state beyond that which can be achieved by filling the pair of primary α4C metal sites alone. Indeed, previous quantitative DNA binding experiments reveal that Cd^II is a relative poorer allosteric negative inhibitor of CmtR^Mtb binding to DNA relative to other ArsR-family metalloregulatory proteins (1, 11); this suggests that Cd^II binding to metal site 2 may further enhance allosteric regulation of DNA binding in a way that leads to increased transcription derepression in the cell. Alternatively or in addition, the second high affinity (K_Cd≥10¹² M⁻¹) cadmium binding site on CmtR^Sc may also simply allow S. coelicolor to sequester (chelate) more Cd^II under conditions of Cd^II stress once CmtR^Sc is dissociated from its DNA operator. Understanding the structural and physiological role of metal site 2 in CmtR^Sc requires further investigation.

CmtR^Sc is not the first ArsR family regulator with two structurally distinct pairs of metalloregulatory sites per dimer. Other examples include the Cd^II/Pb^II sensor pI258 CadC (12) and Cu^I sensor O. brevis BxmR (11). CadC employs four conserved Cys in Cd^II, Co^II and Bi^III binding while adopting an S₃ coordination complex with Pb^II in its primary α3N metal site; Zn^II binding to a pair of C-terminal interhelical α5 sites is functionally silent in CadCs, while other CadCs and α3N-type sensors have dispensed with this site entirely (12, 47). In BxmR, the α3N site is a major sensing site for Cd^II and Cu^I, the latter of which forms a binuclear Cu^I₂S₄ cluster (11); a C-terminal α5 site has evolved exclusively for Zn^II sensing. Thus, in the case of BxmR, the presence of two distinct metal sites relaxes the metal selectivity of this particular sensor. In CmtR^Sc, the impact of metal site 2 on CmtR^Sc function largely parallels the situation in CadCs, where the secondary site does not change the metal specificity profile of the primary metal site (12, 48), but unlike in CadCs, enhances its effectiveness in the cell.

Finally, CmtR^Sc is the first ArsR family metal sensor to be characterized from S. coelicolor and is shown here to possess both a metal sensing site anticipated on the basis of its classification as an CmtR-like α4C Pb^II/Cd^II sensor, as well as a second novel metalloregulatory site not found in other CmtRs. In fact, the presence of consecutive or vicinal Cys residues in what are predicted to be flexible or unstructured N-terminal or C-terminal “tails” like that in metal site 2 in CmtR^Sc, have previously been shown to function as As^III ligands in two arsenic-specific ArsR family proteins recently characterized (49, 50); given this, it seemed possible that Cys110 and/or Cys111 might confer a As^III sensing function on the Pb^II/Cd^II sensing CmtR. Our data reveal that this is clearly not the case, at least as measured by transcriptional derepression in M. smegmatis (Fig. 6). These findings are consistent with the idea that ArsR family sensors may well be readily classified into subfamilies of paralogs that detect distinct cellular inducers solely on the basis of well-characterized primary structural motifs, rather than global sequence similarity (1, 10). However, these core characteristics can be modulated by additional or altogether novel metal sites capable of tuning inducer responsiveness appropriate for the microenvironmental niche in which an organism resides (11, 49).