INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

In eukaryotes, Fe²⁺ and Mn²⁺ are transported by the NRAMP family of transporters. Recently, NRAMP homologues from both gram-negative and gram-positive bacteria have been biochemically characterized as pH-dependent secondary transporters of divalent metal ions with a preference for Mn²⁺, and to a lesser extent, for Fe²⁺ (1, 15, 18, 27), hence the designation MntH, for H⁺-dependent manganese transporter.

Iron and manganese are both toxic at high intracellular concentrations. Therefore, their uptake into bacteria is tightly regulated. In gram-negative bacteria and in gram-positive bacteria with a low GC content, the genes encoding iron uptake proteins are negatively regulated by Fur (ferric uptake regulator). In the presence of Fe²⁺, this protein binds to palindromic sequences within Fur-regulated promoters and represses transcription of the genes. A decreased intracellular iron concentration leads to a lower DNA binding activity of the regulatory protein and concomitantly to derepression of the regulated genes. In gram-positive species with a high GC content, iron regulation is accomplished by the diphtheria toxin repressor (DtxR)-like proteins. Fur and DtxR are representatives of two distinct repressor families. In addition to the iron regulator Fur, the Fur family includes the regulator of Zn²⁺ transport, Zur (8, 11, 21), and the regulator of the peroxide stress response, PerR (4). Recent DNA binding studies have suggested that TroR, a member of the DtxR family, from Treponema pallidum is a Mn²⁺ sensor (25), but it has not been possible to test the role of TroR in vivo. The distantly related MntR from Bacillus subtilis has been described as a bifunctional regulator of two Mn²⁺ transporters. Under low-Mn²⁺ conditions, MntR activates transcription of an ABC transporter, whereas under high-Mn²⁺ conditions, MntR acts as a transcriptional repressor of mntH, an NRAMP homologue (27). Under high-Mn²⁺ conditions, the DtxR homologue ScaR from Streptococcus gordonii represses transcription of the scaABC operon, which encodes the components of an ABC-type transporter for Mn²⁺ (13).

In this work, the regulation of mntH expression in Escherichia coli by divalent metal ions was studied. We report the identification of the gene product of o155, which we renamed MntR, a new Mn²⁺ regulator in E. coli, and the repression of mntH transcription by Mn²⁺-MntR and Fe²⁺-Fur. The MntR protein was purified and shown to bind the mntH operator in vitro.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, plasmids, and growth conditions. M9 minimal medium (19) contained 0.2% glucose as a carbon source. For MA medium, M9 was supplemented with tryptophan, tyrosine, and phenylalanine (0.1 mg ml¹ [each]) and 4-aminobenzoic acid (40 µM) and 4-hydroxybenzoic acid (40 µM). MacConkey plates contained 40 g of MacConkey agar base (Difco, Detroit, Mich.) and 10 g of D-lactose per liter. Mutagenesis with 1-methyl-3-nitro-1-nitrosoguanidine (MNNG), P1 transductions, and -galactosidase assays were performed as described previously (19).

Recombinant DNA techniques. Standard procedures (28) or those recommended by the manufacturer were followed for isolation of chromosomal and plasmid DNA, DNA modification, ligation, transformation, PCR, and agarose gel electrophoresis. DNA was sequenced by the dideoxy chain termination method using an ALFexpress DNA sequencer (Pharmacia Biotech, Freiburg, Germany). Mutations in the chromosomal mntR allele were localized by cycle sequencing of two independently generated PCR-amplified fragments using the primers MNTR1 and MNTR2 and a Thermo Sequenase Cy5 dye terminator kit (Pharmacia Biotech). Oligonucleotides were synthesized by Eurogentec (Seraing, Belgium).

DNase I footprinting. DNase I protection experiments were done as described previously (22). For metal ion binding studies, lower concentrations of MgCl₂ (50 µM) and CaCl₂ (10 µM) were applied. The labeled target DNA was generated by PCR with the primers YFEP9 and carbocyamine dye (Cy5)-labeled YFEP5C (TGTTGTGTATGGAAGCTGAAAG [nt 4581 to 4560, section 217]) for the MntH-coding strand and YFEP3 and fluorescent Cy5-labeled YFEP4C (GCAATGAACGCAGGTCCCATTA [nt 4303 to 4324, section 217]) for the noncoding strand. Standard sequencing reactions were carried out with the Cy5-labeled primer and pSP117/11 as DNA template.

Overproduction and purification of MntR. MntR was overproduced from pSP116/25 using E. coli BL21 (DE3) (31). Bacterial cells were disrupted in a French pressure cell, and the supernatant was chromatographed on a fast protein liquid chromatography MonoQ HR 5/5 column (Pharmacia) with 50 mM Tris-HCl, pH 7.4, and a linear gradient of NaCl (0 to 0.3 M).

Computer analyses. Nucleic acid and amino acid sequences were analyzed by the PC/GENE 6.85 program package (IntelliGenetics, Mountain View, Calif.). For sequence similarity searches, the BLAST facilities of the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov) were used.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Regulation of mntH. To identify new genes regulated by fur in E. coli, random Mud1 phage insertion mutants that were repressed by iron and derepressed by the iron chelator 2,2'-dipyridyl were selected as described in Materials and Methods. The Mud1 insertion site of one mutant (SIP879) was localized in the mntH gene by sequencing the insertion site as described in Materials and Methods. On MacConkey plates, mutant SIP879 formed red colonies, indicating acid production from lactose and high expression of the operon fusion with the lacZ reporter gene. The mntH-lacZ fusion was repressed in cells in the diffusion zone of a filter paper strip impregnated with 100 µM (NH₄)₂Fe(SO₄)₂ or 20 µM MnCl₂, as shown by the formation of white colonies. In cells on MacConkey plates containing 40 µM (NH₄)₂Fe(So₄)₂, the fusion was derepressed by the chelators 2,2'-dipyridyl, desferri-ferrioxamine B, and EDTA (each 10 mM). The -galactosidase activity of mutant SIP879 (mntH-lacZ) was measured in the presence of various divalent metal ions (Fig. 1). Of the metal ions tested, only Fe²⁺ and Mn²⁺ repressed the mntH-lacZ operon fusion fivefold.

View larger version (56K): [in this window] [in a new window]   FIG. 1.   Regulation of mntH-lacZ by divalent metal ions. E. coli SIP879 (mntH-lacZ) was grown aerobically for 6 h in MA medium supplemented with a 10 µM concentration of the indicated metal ion, and -galactosidase activities were then measured. Values are averages of experiments carried out in triplicate.

Iron repression of mntH is mediated by Fur. To determine whether iron regulation is mediated by Fur, the fur gene on the chromosome of strain SIP879 (mntH-lacZ) was inactivated. In the fur mutant SIP890, the mntH-lacZ fusion was no longer repressed by iron, whereas regulation by manganese was unimpaired (Fig. 2A). The isogenic fur⁺ strain SIP891 showed the same regulation as strain SIP879. The derepression in the presence of iron in the fur mutant was complemented by the fur-carrying plasmid pSP61/18 (Fig. 2A). This indicates that Fur alone is responsible for the iron repression. The same results were obtained for the aroB⁺ strains SIP894 (mntH-lacZ fur) and SIP895 (mntH-lacZ) (data not shown).

View larger version (45K): [in this window] [in a new window]   FIG. 2.   Response of mntH-lacZ in wild-type, mntR mutant, and fur mutant cells to various concentrations of (NH4)2Fe(SO4)2 (A) and MnCl2 (B). The mntR mutant SIP933 was complemented with pSP116/1 (plasmid carrying mntR), and the fur mutant SIP890 was complemented with pSP61/18 (plasmid carrying fur). Cells were grown aerobically for 6 h in MA medium (white bars) or in MA medium amended with 5 µM (black bars) or 20 µM (gray bars) metal ions, and -galactosidase activities were then determined. Values are averages of experiments carried out in triplicate.

Localization of the Fur binding site in vivo. To localize the DNA region within the mntH operator responsible for Fur binding, an in vivo titration assay (30) was employed. Introduction of a Fur binding site on a high-copy-number plasmid titrates Fur and hence derepressed the Fur-regulated fhuF'-lacZ expression in E. coli H1717, resulting in red colonies on MacConkey plates containing 40 µM (NH₄)₂Fe(SO₄)₂. In contrast, E. coli H1717 harboring the vector forms white colonies. E. coli H1717 transformed with the plasmid pSP117/11 carrying the mntH operator region (nt 3610 to 4764, section 217) cloned on pBC-SK⁺ resulted in red colonies (Fig. 3), indicating Fur binding, whereas the colonies of the vector control were white. The region responsible for the Fur regulation was narrowed down to nt 4433 to 4455 of section 217 by subcloning as depicted in Fig. 3. Deletion of nt 4452 from this region (pSP118/14) resulted in lower activity, whereas pSP118/18, comprising nt 4433 to 4454, was fully active in the in vivo titration assay (data not shown). The binding site GAgAATGATtATCAaatTC matches the Fur-box consensus sequence GATAATGATAATCATTATC (9) in 14 of 19 nt (mismatches are shown as lowercase letters).

View larger version (17K): [in this window] [in a new window]   FIG. 3.   Mapping of the Fur box and of the potential MntR binding site on the mntH operator region. The DNA inserts (shaded boxes) were obtained with the primers specified in Materials and Methods and were then ligated into the EcoRV site of pBC-SK+. Thick arrows show the position and orientation of the lacZ promoter in each construct. The numbers reflect the positions in section 217 of the E. coli genome (3). Activity in the Fur titration assay signifies derepression of fhuF'-lacZ expression in E. coli H1717 carrying the corresponding plasmid. This results in red colonies on MacConkey plates containing 40 µM (NH4)2Fe(SO4)2. E. coli H1717(pBC-SK+) cells grew as pale colonies on these plates. For the MntR titration assay, E. coli SIP879 was transformed with the plasmids and analyzed on MacConkey plates with 80 µM MnCl2. E. coli SIP879(pBC-SK+) formed white colonies; colonies of transformants with activity were reddish. Parentheses indicate partially reduced activity. Plus and minus signs indicate the relative amount of activity, with ++ being the highest relative amount.

Identification of the Mn²⁺ regulator MntR. To find the regulator responsible for the Mn²⁺ regulation, strain SIP879 (mntH-lacZ) was mutagenized with MNNG and screened for derepressed red colonies on MacConkey plates containing 10 µM Mn²⁺. The impaired manganese repression of the selected mutants is summarized in Table 2. Mutant SIP932 was transformed with an E. coli gene library (21) and screened for white colonies on MacConkey manganese plates. One plasmid (pSP115/25) complemented strain SIP932 as well as mutants SIP924, SIP931, SIP933, SIP943, and SIP949. The plasmid comprises nt 10296 of section 73 to nt 1924 of section 74 of the E. coli genome. Subclones were constructed, of which plasmids pSP116/1 and pSP116/10 contained only o155 and restored repression by manganese (see Fig. 2B for pSP116/1). The mutations were cotransducible with the tetracycline resistance marker zbh-272::Tn10 at 18.4 min on the genetic map of E. coli. This confirmed that the Mn²⁺-deregulated mutants are mutated in the region of o155. By sequencing, the mutations were all localized in o155 (Table 2). The amino acid sequence of O155 reveals 16% identity to MntR from B. subtilis and shows that it belongs to the DtxR family of bacterial metalloregulators. Hence o155 was renamed mntR (Mn²⁺ transporter regulation).

Subcloning of the mntH operator region affording manganese regulation in vivo. When cloned on the high-copy-number vector pBC-SK⁺, the region nt 3610 to 4764 of section 217 covering the mntH operator (pSP117/11) titrated MntR from the chromosomal mntH-lacZ fusion of strain SIP879; colonies of SIP879 transformed with pSP117/11 were reddish on MacConkey plates containing 80 µM MnCl₂, whereas SIP879 transformed with the vector pBC-SK⁺ formed white colonies. This indicates that the insert binds to MntR. By subcloning, this MntR binding site was narrowed down to nt 4391 to 4440 of section 217 (Fig. 3), which contains a nearly perfect palindrome (see Fig. 5C).

Interaction of MntR with the mntH operator region. After overproduction, MntR was purified by anion-exchange chromatography to electrophoretic homogeneity. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of purified MntR revealed a single protein with a molecular mass of 17 kDa, consistent with the predicted size of 17.6 kDa (Fig. 4). The DNA binding properties of MntR were investigated by DNase I footprint assays. The target DNA fragment encompassed at least 195 nt upstream of the translational start of mntH and was labeled on each strand. Footprinting experiments in the presence of Mn²⁺ showed a core region of protection that is depicted in Fig. 5. At higher MntR concentrations, partial protection adjacent to the core region occurred. Even without added Mn²⁺ the DNA was protected, probably due to the high Mg²⁺ concentrations (5 mM) in the DNase I buffer. Therefore, lower Mg²⁺ concentrations at which DNase I was still active were used for metal ion specificity studies (see Materials and Methods). The low-specificity ion chelator EDTA (100 µM) impaired DNA protection. Addition of 150 µM Mn²⁺ restored DNA binding, whereas no protection was observed in the presence of 150 µM Co²⁺, Cu²⁺, Fe²⁺, Ni²⁺, or Zn²⁺. Zn²⁺ did not protect but altered the DNase I fragmentation pattern.

View larger version (30K): [in this window] [in a new window]   FIG. 4.   Sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis of the purified MntR. mntR was overexpressed using the T7 expression system on pSP116/25 carried by strain BL21(DE3). Lane 1, soluble fraction of cell lysate after induction; lane 2, MntR purified by MonoQ anion-exchange chromatography. Proteins were stained with Coomassie blue. The positions and molecular masses of the standard proteins are shown on the left.

View larger version (28K): [in this window] [in a new window]   FIG. 5.   DNase I footprinting analysis of the MntH-coding (A) and noncoding (B) strands of the mntH operator with MntR protein. The 311-bp DNA fragment comprising nt 4581 to 4271 of section 217, labeled on the MntH-coding strand with the fluorescence-labeled YFEP5C primer (A), and the 462-bp fragment encompassing nt 4303 to 4764 of section 217, labeled on the noncoding strand with the Cy5-labeled YFEP4C primer (B), were incubated without (row 1) or with (row 2) purified MntR protein as described in Materials and Methods. The nucleotide sequence obtained from the dideoxynucleotide sequencing reaction with the same end-labeled primer is given at the bottom of each panel; the protected region is expanded for clarity. (C) Nucleotide sequence of the mntH operator region with the translational start site. The boxed region indicates the nucleotides protected from DNase I by MntR. The area active in the in vivo Fur titration assay is shaded. Palindromic sequences are denoted by convergent arrows, with complementary bases shown in bold. Nucleotides are numbered according to E. coli genome section 217.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

This study identifies a new metalloregulatory protein in E. coli, MntR, that senses Mn²⁺ and belongs to the DtxR family. The crystal structures of two iron regulatory proteins of the DtxR family, DtxR from Corynebacterium diphtheriae and IdeR from Mycobacterium tuberculosis, have been solved (24, 26). The proteins comprise three distinct domains: an amino-terminal DNA binding domain containing a helix-turn-helix motif, a central dimerization domain including the key residues for metal ion coordination, and a carboxy-terminal -spectrin SH3-like domain proposed to regulate repressor activity and also to contribute to metal ion binding (23). In contrast to the Mn²⁺-responsive metalloregulator ScaR (13), MntR from E. coli, the Mn²⁺-sensing TroR from T. pallidum (25), and MntR from B. subtilis (27) lack the C-terminal third (SH3-like) domain of DtxR. Therefore, the third domain is not correlated with the metal ion specificity for Mn²⁺ or Fe²⁺.

Regulation of the transporter gene mntH by Mn²⁺ is only accomplished by MntR and not by Fur, although the E. coli Fur protein also functions in vitro with Mn²⁺ as a corepressor. This might also occur in vivo under certain conditions. At high, toxic Mn²⁺ concentrations, iron uptake is repressed and growth ceases. Mutants selected for Mn²⁺ resistance are often mutated in fur, which allows the cells to synthesize specific iron uptake systems (12).

It seems reasonable that in E. coli the NRAMP homologue transporter MntH, which takes up Mn²⁺ as well as Fe²⁺, is regulated by both ions via specific regulators. In contrast to mntH from E. coli, no iron regulation has been reported for mntH from B. subtilis (27). It is possible that in B. subtilis the repression by Fe²⁺ is only found at iron concentrations higher than the ones tested (up to 1 µM). Besides the NRAMP homologous proton-coupled Mn²⁺ transporter MntH, B. subtilis contains the ABC-type Mn²⁺ transporter MntABCD, which is specific for Mn²⁺ and which is activated by MntR under low-Mn²⁺ conditions (27). ScaR from S. gordonii and TroR from T. pallidum have been described as repressors of ABC-type transporters, probably for Mn²⁺ uptake (13, 25). The ScaABC transporter from S. gordonii is not controlled by iron (13). It is unlikely that E. coli also possesses an ABC transporter for Mn²⁺ uptake since Mn²⁺ uptake is reported to be solely dependent on the membrane potential rather than on ATP (2, 29).

MntR protected a core region of 25 nt on the coding strand and on the noncoding strand of the mntH operator with a 3' stagger. This is in accordance with other classic helix-turn-helix repressors that bind to DNA as a dimer, e.g.,  phage repressor CI (14, 34). Higher concentrations of MntR extended the protection zone in DNase I footprinting assays. Similarly, for ScaR and TroR, which occupy 22 nt with a palindromic sequence, at least two distinct DNA-protein complexes have been observed in mobility-shift DNA binding assays, which indicates multiple binding interactions between the regulator and the operator (13, 25). Likewise, polymerization of the Fur repressor has been observed in footprinting experiments (e.g., see reference 9) and by electron and atomic force microscopy (10, 16); however, the crystal structure of the DNA complex is not yet known. The structure of the DtxR-DNA complex revealed that DNA surprisingly interacts with two dimeric repressor proteins bound to opposite sides of the operator (35). Similar to ScaR, TroR, and DtxR, MntR-DNA binding seems to be more complex than that of a classic repressor. Consistent with this complexity, MntR from B. subtilis acts not only as a repressor but also as an activator (27).

The sequence within the promoter recognition region of mntH shows similarity to the sequences bound by the MntR regulator of B. subtilis (27) (Fig. 6A) and low similarity to the Mn²⁺-responsive metalloregulators, TroR from T. pallidum and ScaR (13) (Fig. 6B), to the DtxR consensus sequence (17), and to the Fur-box consensus sequence (9). The operator region of E. coli mntR contains no recognizable MntR box, giving no hint for autoregulation, which has not been examined further.

View larger version (15K): [in this window] [in a new window]   FIG. 6.   Analysis of the manganese repressor binding site sequences. (A) Alignment of the MntR box from E. coli with the putative MntR recognition site from the mntH and mntABCD control regions of B. subtilis (27). (B) Comparison of the MntR box from E. coli with the binding site for TroR from T. pallidum (25) and ScaR from S. gordonii (13). Bases that are identical in at least two sequences are shaded; bases common to all three sequences are marked by asterisks.

In many enzymes, Mn²⁺ can be replaced by Mg²⁺ and vice versa. For example, in the DNase I footprinting assays in this study, DNase I activity was strongly enhanced by Mn²⁺ in the absence of Mg²⁺. Thus, the physiological relevance of manganese for enzyme activity is often not known. However, in E. coli, a number of metalloenzymes that require Mn²⁺ are known, e.g., Mn²⁺-containing superoxide dismutase (MnSOD), protein phosphatases PrpA and -B, cofactor-independent 3-phosphoglycerate mutase (iPGM), agmatinase, phosphoenolpyruvate carboxylase, and exonuclease SbcCD. Possibly because there are enzymes that certainly need Mn²⁺ as a cofactor, the MntH uptake system is not completely repressed by Fe²⁺-Fur (Fig. 1 and 2). Derepression of the Mn²⁺ transport system via the Fur system may indicate that Mn²⁺ or Mn²⁺-containing enzymes substitute for iron or iron-containing enzymes under iron limitation.