Genes Encoding Specific Nickel Transport Systems Flank the Chromosomal Urease Locus of Pathogenic Yersiniae

Abstract

The transition metal nickel is an essential cofactor for a number of bacterial enzymes, one of which is urease. Prior to its incorporation into metalloenzyme active sites, nickel must be imported into the cell. Here, we report identification of two loci corresponding to nickel-specific transport systems in the gram-negative, ureolytic bacterium Yersinia pseudotuberculosis. The loci are located on each side of the chromosomal urease gene cluster ureABCEFGD and have the same orientation as the latter. The yntABCDE locus upstream of the ure genes encodes five predicted products with sequence homology to ATP-binding cassette nickel permeases present in several gram-negative bacteria. The ureH gene, located downstream of ure, encodes a single-component carrier which displays homology to polypeptides of the nickel-cobalt transporter family. Transporters with homology to these two classes are also present (again in proximity to the urease locus) in the other two pathogenic yersiniae, Y. pestis and Y. enterocolitica. An Escherichia coli nikA insertion mutant recovered nickel uptake ability following heterologous complementation with either the ynt or the ureH plasmid-borne gene of Y. pseudotuberculosis, demonstrating that each carrier is necessary and sufficient for nickel transport. Deletion of ynt in Y. pseudotuberculosis almost completely abolished bacterial urease activity, whereas deletion of ureH had no effect. Nevertheless, rates of nickel transport were significantly altered in both ynt and ureH mutants. Furthermore, the ynt ureH double mutant was totally devoid of nickel uptake ability, thus indicating that Ynt and UreH constitute the only routes for nickel entry. Both Ynt and UreH show selectivity for Ni²⁺ ions. This is the first reported identification of genes coding for both kinds of nickel-specific permeases situated adjacent to the urease gene cluster in the genome of a microorganism.

Urease, produced by a wide range of eukaryotic and prokaryotic organisms, is an enzyme that hydrolyzes urea to give ammonia and carbamate. This multimeric enzyme requires nickel for activity and is thus classed as a metalloenzyme (23). Bacteria have devised elaborate mechanisms for acquisition and incorporation of the divalent nickel ion. Nickel uptake constitutes the first step in this process. Diffusion of cations through the outer membrane of gram-negative bacteria occurs preferentially via the OmpC and OmpF porins (25). Translocation of nickel through the cytoplasmic membrane is then mediated by the nonspecific CorA and Mgt magnesium transport systems (33) and/or high-affinity, nickel-specific permeases (12). To date, two classes of high-affinity nickel transporters have been described for bacteria. The first, the Nik system, was originally identified as being essential for hydrogenase activity in Escherichia coli and is a member of the ATP-binding cassette (ABC) family. It consists of five components: a periplasmic binding protein (NikA) exhibiting high specificity for nickel, two integral cytoplasmic membrane proteins (NikB and NikC) assumed to form a channel for nickel transport, and finally, two membrane-associated ATPases (NikD and NikE) which are responsible for coupling energy to the transport process (24). Expression of the E. coli nik operon is activated (in the absence of oxygen) by the global regulatory protein FNR and repressed (in the presence of high nickel concentrations) by the nickel-responsive regulator NikR, which is functionally similar to the Fur ferric ion uptake regulator (5, 8, 39). Similar ABC transport systems with significant homology to the E. coli Nik proteins have recently been described for two human pathogens, Vibrio parahaemolyticus (26) and Brucella suis (18).

The second type of nickel importer is a single-component permease; the prototype of this class is HoxN from Ralstonia eutropha (previously Alcaligenes eutrophus). HoxN is an integral protein containing eight membrane-spanning segments (13). Members of this family have been identified in a number of bacteria, including Helicobacter pylori (22), Bradyrhizobium japonicum (14), and Mycobacterium tuberculosis (6), as well as in the fission yeast, Schizosaccharomyces pombe (11). Mutations in either specific transport system lead to a dramatic reduction in bacterial nickel uptake and, consequently, to decreased activity of the relevant nickel-requiring urease and hydrogenase enzymes (12, 17, 24).

The gram-negative enteropathogenic bacterium Yersinia pseudotuberculosis is a ureolytic species responsible for self-limiting, intestinal tract infections in humans. Genes involved in urease biosynthesis have recently been characterized (28). Three adjacent chromosomal genes (ureA, ureB, and ureC) encode the structural subunits which associate to constitute an inactive apoenzyme. Incorporation of nickel ions into the enzyme's catalytic site (located in the UreC protein) requires at least four additional genes (ureE, ureF, ureG, and ureD, situated in that order on the chromosome) contiguous to the structural genes. In this report, we show that Y. pseudotuberculosis produces two different types of specific nickel transporters (the synthesis of which is mediated by genes flanking the 5′ and 3′ ends of the ure locus): a multicomponent ABC nickel transporter encoded by the yntABCDE locus and located upstream of ure and a single-component transporter encoded by the ureH gene and located downstream of ure.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

The main characteristics of the bacterial strains and plasmids used in this study are listed in Table 1. Y. pseudotuberculosis and E. coli strains were grown at 28 and 37°C, respectively, in Luria-Bertani (LB) broth or on agar plates. Mating experiments were carried out by plating on M9 minimum medium agar (33 mM Na₂HPO₄, 22 mM KH₂PO₄, 8.5 mM NaCl, 18 mM NH₄Cl, 1 mM MgSO₄, 0.1 mM CaCl₂, 0.3 μM thiamine, 10 mM glucose, 14 g of agar/liter). Ampicillin (100 μg ml⁻¹), kanamycin (50 μg ml⁻¹), and chloramphenicol (50 μg ml⁻¹) were added to growth media for bacterial selection when necessary.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Relevant property(ies)a	Reference or origin
Strains
Yersinia pseudotuberculosis
32777	Wild-type strain	28
MYUH	Derived from strain 32777, ΔureHΩaphA-1a	This work
MYNT	Derived from strain 32777, ΔyntABCDEΩcat	This work
MYOU	Derived from strain 32777, ΔyntABCDEΩcat ΔureHΩaphA-1a	This work
Escherichia coli
DH5α	supE ΔlacU169 (φ80dlacZΔM15) thi deoR phoA hsdR recA endA gyrA relA	16
SY327λpir	Δ(lac pro) argE(Am) recA rif nalA λpir; host for pCVD442 and derivatives	21
SM10λpir	thi thr leu sup tonA lacY recA::RP4-2Tc::MuKm λpir; host for pCVD442 and derivatives	32
P4X	Hfr metB	24
HPX72	Derived from strain P4X, nikA::MudI (Km lac)	24
Plasmids
pZero2-1	Cloning vector, Km	Invitrogen
pUC18	Cloning vector, Ap	36
pACYC184	Plasmid vector, Cm Tc; source of the chloramphenicol acetyltransferase cat gene	4
pUC4K	Plasmid vector; Km; source of the aminoglycoside phosphotransferase aphA-1a gene	34
pCVD442	Suicide vector containing the counterselectable marker sacB; Ap	9
pMS89	pHC79Ω, ∼35-kb Sau3A DNA fragment from Y. pseudotuberculosis 32777 containing the urease locus and its flanking regions	28
pLW21	pUC19Ω, 7.1-kb SmaI insert with the nikABCDER region from E. coli K-12	24
pFS45	pUC18Ω, ∼2.4-kb XhoI/Bst1107I insert with the 3′ end of yut and ureH from Y. pseudotuberculosis 32777	This work
pFS50	pACYC184Ω, ∼7.3-kb HindIII/PstI insert with the Y. pseudotuberculosis 32777 ure locus and its promoter region	This work
pFS78	pUC18Ω, ∼5.5-kb PCR-generated fragment with primer set N1-N13 containing ynt genes from Y. pseudotuberculosis 32777	This work
pFSU1	pZero2-1Ω, ∼1.2-kb XbaI/EcoRI PCR-generated fragment with primer set N5-N6 encompassing the upstream region of ureH gene from Y. pseudotuberculosis 32777	This work
pFSU2	pFSU1Ω, ∼0.8-kb SacI/EcoRI PCR-generated fragment with primer set N7-N8 encompassing the downstream region of ureH gene from Y. pseudotuberculosis 32777	This work
pFSU3	pFSU2Ω, ∼1.3-kb EcoRI PCR-generated fragment with primer set K1-K2 encompassing aphA-1a from pUC4K	This work
pFSΔU	pCVD442Ω, ∼3.3-kb XbaI/SacI insert from pFSU3	This work
pFSO1	pUC18Ω, ∼0.5-kb HindIII/PstI PCR-generated fragment with primer set N1-N2 encompassing the upstream region of yntA gene from Y. pseudotuberculosis 32777	This work
pFSO2	pFSO1Ω, ∼0.5-kb BamHI/EcoRI PCR-generated fragment with primer set N3-N4 encompassing the downstream region of yntE gene from Y. pseudotuberculosis 32777	This work
pFSO3	pFSO2Ω, ∼1.4-kb PstI/BamHI PCR-generated fragment with primer set C3-C4 encompassing the cat gene from pACYC184	This work
pFSΔO	pCVD442Ω, ∼2.4-kb SacI insert from pFSO3	This work

^aAp, Cm, Km, and Tc, resistance to ampicillin, chloramphenicol, kanamycin, and tetracycline, respectively. Ω, in vitro insertion.

Yersinia and Escherichia strains were characterized with API 20E strips (bioMérieux). Detection of urease activity was performed with urea agar (31) and urea-indole medium (Diagnostics Pasteur).

Nucleic acid manipulations.

Genomic DNA extraction and small-scale isolation of plasmid DNA were performed as previously described (29). Large-scale plasmid DNA preparations were purified on Qiagen columns in accordance with the manufacturer's recommendations (Qiagen GmbH). Genomic and plasmid DNAs were digested with the appropriate restriction endonucleases purchased from GIBCO BRL or Promega: the resulting fragments were separated by agarose gel electrophoresis (0.8 to 1.2% agarose) and transferred onto Hybond-N⁺ membranes (Amersham) by the Southern technique. Restriction fragments were eluted from agarose gels with the Qiaquick gel extraction kit (Qiagen GmbH). DNA fragments were ligated to endonuclease-restricted vectors via standard T4 DNA ligase procedures (GIBCO BRL). Recombinant plasmid DNAs were introduced into E. coli and Yersinia by transformation (29) and electroporation (7), respectively.

Prehybridization, hybridization of membrane-blotted DNA or RNA with digoxigenin-labeled DNA probes under stringent conditions, and detection of nucleic acid hybrids were all performed with the DIG hybridization and detection kit from Boehringer Mannheim.

Nucleotide sequence determination was performed by the dideoxy chain-termination method, with the ABI PRISM dichloRhodamine dye terminator sequencing kit with Amplitaq DNA polymerase FS (Perkin-Elmer), according to the manufacturer's instructions. Extension products were analyzed with the Applied Biosystems ABI 3700 automated DNA sequencer (Perkin-Elmer). Nucleotide sequences were analyzed with Perkin-Elmer software (Sequence Analysis and Sequence Navigator). Multiple protein alignment was carried out with the CLUSTAL_X program.

PCRs.

PCR amplification was performed in a 100-μl reaction volume with a thermal cycler (the 2400 model from Perkin-Elmer Cetus). Fifty nanograms of target DNA, 0.1 nmol of each primer, and 1 U of thermostable DNA polymerase were mixed in the corresponding 1× polymerase buffer (200 mM in each deoxynucleotide triphosphate). Amplification involved 30 cycles, each consisting of (i) a 1-min denaturation step at 94°C, (ii) a 1-min annealing step at 55°C, and (iii) a 1-min polymerization step at 72°C. Digoxigenin-labeled PCR products were generated with PCR DIG labeling mix from Boehringer Mannheim. Amplimers were purified on SpinX columns (Corning Costar Corporation).

Oligonucleotide primers.

Nineteen primers were synthesized (by Sigma and Genset) for PCR generation of DNA fragments to be cloned or used as probes. The 5′→3′ nucleotide sequences were as follows: N1, CCCAAGCTTGAGCTCGCCTGATGCCTTTGGTGTGT; N2, TGCACTGCAGAGTGATTGCCTGTCAGGCA; N3, CGGGATCCCATTTGGGGTTAGCAATGG; N4, CCGGAATTCGAGCTCCTGACGCAGCATTTACCATC; N5, TATCAGCCAGTGATCCAAGCA; N6, TCAACCCTACCCGTTCTGAC; N7, GCGGAATTCGAGTGAATTATTAGACCCGC; N8, GTCGAGCTCGATGCAATCCAACATATCGC; N9, GGTTGTAATACGCATGAGCC; N10, CTCAGTGAGCGAATTCAAC; N11, CGTGGTTGCTGACACTTAAG; N12, CGCCAATTATTGGCCGATCA; N13, ATCTGGCCAATAACTGCG; K1, GCTCTGAATTCGATATCGGGGAAAGCCACGTTGTGTC; K2, GATTGGAATTCGATATCCTGAGGTCTGCCTCGTGAAGAA; C1, TCAGCGCTAGCGGAGTG; C2, GATCTGCATCGCAGGAT; C3, TGCACTGCAGCACTCCGCTAGCGCTGA; and C4, CGGGATCCGATCTGCATCGCAGGAT.

Urease extract preparation and enzyme activity measurement.

Overnight Yersinia cultures at 28°C in LB broth were adjusted to 10⁵ cells per ml; 50 μl of bacterial culture was then added to 50 ml of fresh LB broth. Yersinia bacteria from a 36-h (stationary-phase) culture were harvested by centrifugation (2,600 × g) for 5 min at 4°C and washed twice with 0.2 M sodium phosphate buffer (pH 6.8). Bacterial cells were resuspended in phosphate buffer and disrupted twice with a French press (10,000 lb/in²). Following centrifugation (12,800 × g) for 30 min at 4°C, supernatants were placed on ice. Protein concentration was determined by the Bradford dye-binding procedure according to the manufacturer's instructions (Bio-Rad). The urease activity of extracts was determined by measuring the amount of ammonia released from urea in the phenol-hypochlorite assay (20). Two micrograms of total proteins from bacterial extracts was added to 200 μl of 50 mM urea in 0.1 M sodium phosphate (pH 6.8), and the mixture was incubated at 37°C for 20 min. The reaction was stopped by addition of 400 μl of phenol-nitroprusside solution (50 g of phenol and 250 mg of sodium nitroprusside per liter). Four hundred microliters of sodium hydroxide (11 N)-sodium hypochlorite (0.175% [vol/vol]) solution was added, and the contents were mixed well. Following incubation at 50°C for 6 min, the absorbance at 625 nm was measured. A standard ammonium chloride concentration curve was determined to be linear between 28 and 448 nmol of ammonia. Absorbance values were converted to nanomoles of ammonia based on the ammonium chloride standard curve. Data are presented in terms of urease specific activity, defined as micromoles of NH₃ per minute per milligram of protein. Two micrograms of total proteins from bacterial extracts boiled for 5 min served as the negative control, and the background was subtracted from all values obtained to avoid measuring ammonia generated by urease-independent reactions.

Nickel transport.

E. coli and Y. pseudotuberculosis strains were grown until the mid-exponential growth phase under microaerobic conditions in LB broth medium supplemented with molybdate and selenite, as previously described (39). The cells were washed twice and resuspended in 1 ml of transport buffer (66 mM KH₂PO₄-K₂HPO₄ [pH 6.8], 11 mM glucose, 10 mM MgCl₂, 0.5 mM dithionite) to a final concentration of approximately 3 to 5 mg of dry matter/ml. Solutions were purged with nitrogen and equilibrated in a 30°C water bath for 5 min.

The assay was initiated by the addition of 25 to 150 nM ⁶³NiCl₂ (0.92 mCi μmol⁻¹; Amersham). The cation specificity was examined by using 5 μM CdCl₂, CoSO₄, CuSO₄, MnSO₄, or ZnSO₄ in the presence of 150 nM ⁶³NiCl₂. Samples (0.1 ml) were taken at regular time intervals, filtered through cellulose nitrate membrane filters (Whatman; 0.45-μm pore size), and washed twice with 2 ml of rinsing buffer (66 mM KH₂PO₄-K₂HPO₄ [pH 6.8], 10 mM EDTA). Filters were placed in scintillation vials containing 5 ml of scintillation fluid (ACS; Amersham) for counting in a Packard liquid scintillation counter. Nickel uptake is expressed as picomoles of Ni²⁺ taken up per milligram (dry weight) of bacteria.

Nucleotide sequence accession number.

The nucleotide sequence data reported here have been deposited in the GenBank nucleotide sequence database (accession no. AF412327 and AF412328).

RESULTS

ureH, encoding a putative nickel transporter, is located downstream of the Y. pseudotuberculosis urease locus.

We sequenced the 3′ region flanking the ure locus by using cosmid pMS89, which contains a ∼30-kb DNA fragment encompassing the ureABCEFGD urease-encoding genes of Y. pseudotuberculosis 32777 (28). We identified a gene encoding a urea transporter (referred to as yut) (30) adjacent to and 378 nucleotides downstream of ureD, the last gene in the ure locus. Subsequently, we identified another gene (ureH, 1,059 bp) situated 148 bp downstream of yut: ureH codes for a putative 353-amino-acid protein (calculated mass, 38,693 Da) homologous to nickel permeases from R. eutropha (HoxN; 45 and 59% identity and similarity, respectively), B. japonicum (HupN; 43 and 57% identity and similarity, respectively), M. tuberculosis (NicT; 37 and 52% identity and similarity, respectively), and H. pylori (NixA; 35 and 51% identity and similarity, respectively) (Fig. 1). The similarities between the putative product of ureH and these proteins, together with the ureH gene's location close to the Y. pseudotuberculosis ure locus, strongly suggested that UreH is a nickel transporter. Computerized hydropathy analysis of UreH (data not shown) indicated that the protein may form eight transmembrane segments, with the carboxy- and amino-terminal regions localized in the cytoplasmic compartment—a topological feature shared by several single-peptide nickel transporters (13, 15). Four motifs which are critical for transport activity in single-component nickel permeases (13, 15) were found in UreH at positions 61 to 69, 93 to 101, 207 to 214, and 244 to 252, respectively (Fig. 1).

FIG. 1.

Alignment of the amino acid sequence of the UreH protein from Y. pseudotuberculosis 32777 with those of nickel transporters from R. eutropha (HoxN), B. japonicum (HupN), M. tuberculosis (NicT), and H. pylori (NixA). Asterisks, colons, and periods indicate identical, similar, and related amino acids, respectively. Dashes correspond to gaps introduced to optimize homology between sequences. Motifs critical for biological activity are overlined.

Inactivation of the ureH gene does not abolish urease activity in Y. pseudotuberculosis.

To test whether the UreH nickel carrier is essential for urease activity in Y. pseudotuberculosis, a ureH mutant (MYUH) was made in wild-type strain 32777. A 1,093-bp intragenic deletion was replaced by a kanamycin resistance cassette. This was introduced by allelic exchange after mating Y. pseudotuberculosis strain 32777 with E. coli strain SM10 λpir harboring the pFSΔU plasmid (Table 1). The MYUH ureH mutant was selected on sucrose agar (3). Its genotype was confirmed by PCR assays with primer sets N11-K2 and N12-K1 and by Southern blot hybridization with two appropriate DNA probes, one corresponding to the kan cassette and the other corresponding to the upstream region of ureH (Fig. 2).

FIG. 2.

Genetic organization of the immediate environment of the chromosomal urease (ure) locus and genotype analysis of ynt- and ureH-deficient mutants of Y. pseudotuberculosis 32777. (A) KpnI (K) and EcoRI (E) restriction map of the chromosome of wild-type strain 32777 and isogenic mutants MYUH (ureH) and MYNT (ynt). (B) (Left) Southern blot of KpnI-digested DNA from wild-type strain 32777 and the yntABCDE-deficient mutant hybridized with probe 1 (0.5 kb, corresponding to the downstream region of yntE) and probe 2 (1.4 kb, detecting the chloramphenicol resistance gene cat). (Right) Southern blot of EcoRI-digested DNA from wild-type strain 32777 and the ureH-deficient mutant hybridized with probe 3 (0.6 kb, corresponding to the upstream region of ureH) and probe 4 (1.3 kb, detecting the kanamycin resistance gene aphA-1a). Primers N1 to N12, K1, K2, C1, and C2 were used both for creating mutants and for checking their genotype. Numbers in parentheses indicate the DNA sequence coordinates.

The ability of the MYUH mutant to degrade urea was tested in bacteria cultured in LB broth. Surprisingly, no significant difference in ureolytic activity was found compared with the wild-type strain. Enzyme activity was determined by using media with limiting and high nickel ion concentrations in order to account for the possibility that UreH functionality might be metal dose dependent. No difference (variance analysis: F test, P > 0.05) between the wild-type and ureH mutant strains was detected after growth of bacteria in LB broth containing increasing amounts of either the nickel chelator nitrilotriacetic acid or nickel chloride (Fig. 3). Furthermore, urease activity of the mutant was similar to that of the parent (data not shown) when the bacteria were cultured in the presence of 100 mM MgCl₂, which inhibits nonspecific, divalent cation transporters (11) that could otherwise impede detection of nickel transport by UreH.

FIG. 3.

Ureolytic activity of wild-type strain 32777 (gray bars) and the MYUH ureH mutant (black bars) following growth in LB broth containing increasing concentrations of the nickel chelator nitrilotriacetic acid (NTA) (A) or NiCl₂ (B). Each bar is the mean value of three independent experiments ± the standard deviation.

Nickel uptake is restored in an E. coli nikA mutant trans-complemented with the Y. pseudotuberculosis ureH gene.

Although no defect in ureolytic activity could be observed in the ureH mutant, UreH was shown to behave as a nickel carrier by heterologous complementation of an E. coli nickel transporter-deficient nikA mutant (HPX72) with the Y. pseudotuberculosis ureH gene. Like almost all E. coli strains, strain HPX72 is nonureolytic. When this strain was cotransformed with compatible plasmids pFS45 bearing the ureH gene and pFS50 harboring the Y. pseudotuberculosis ure locus with its promoter region, it was able to degrade urea. As a control, strain HPX72 containing pFS50 alone was nonureolytic. The presence of the ureH gene restored hydrogenase activity to the same level as that of the E. coli wild-type strain P4X (data not shown). Therefore, the Y. pseudotuberculosis ureH gene appears to be able to functionally replace the E. coli nik locus, thus allowing production of active urease and hydrogenase. Secondly, nickel uptake assays were performed in order to substantiate the role of UreH in nickel transport. Assays were conducted in the presence of 10 mM MgCl₂ and at low nickel concentrations (150 nM). Introduction of plasmid pFS45 restored HPX72's rate of nickel transport to a level similar to that obtained by homologous complementation of the mutant with pLW21 bearing the entire E. coli nik locus (Fig. 4A).

FIG. 4.

Roles of UreH and Ynt in nickel entry into bacteria. Bacterial suspensions were incubated in the presence of 150 nM ⁶³NiCl₂ and 10 mM MgCl₂. Nickel uptake was assessed at regular intervals. Three separate experiments gave similar results. Representative data are shown. (A) Nickel uptake into nikA-deficient E. coli strain trans-complemented with ureH (pFS45) or yntABCDE (pFS78) genes from Y. pseudotuberculosis and nikABCDE genes (pLW21) from E. coli. (B) Nickel uptake into Y. pseudotuberculosis ureH, yntABCDE, and ureH yntABCDE mutants derived from the wild-type strain.

The 5′ end of the Y. pseudotuberculosis urease locus is flanked by the Ynt ABC-nickel transport system.

Since urease activity in Y. pseudotuberculosis was not impaired by ureH inactivation, we suspected the existence of another nickel carrier in this bacterium. Sequencing of the 5-kb DNA region upstream of ureA allowed identification of five open reading frames (ORFs). ORF1 (1,578 bp), ORF2 (969 bp), ORF3 (807 bp), ORF4 (807 bp), and ORF5 (666 bp) are apparently arranged in an operon-like manner since (i) all are transcribed in the same direction and (ii) intergenic spaces between neighboring ORFs vary between −1 and −8 nucleotides. ORF5 was situated 1,198 bp from ureA, the first gene of the ure locus. The putative proteins encoded by these five ORFs show striking amino acid sequence similarities to components of the E. coli Nik specific nickel ABC transport system—between 25 and 32% identity, depending on the operon component in question (24). ORF1 encodes a 525-amino-acid protein (predicted molecular mass of 57,820 Da) which is homologous to the E. coli NikA periplasmic nickel-binding protein. The putative ORF2 and ORF3 polypeptides (322 and 268 amino acids with calculated molecular masses of 35,233 and 30,002 Da, respectively) were predicted to contain six potential transmembrane domains, displaying similarity to the NikB and NikC integral cytoplasmic membrane proteins. Like NikD and NikE, ORF4 (268 amino acids, molecular mass of 29,472 Da) and ORF5 (221 amino acids, molecular mass of 24,754 Da) possess characteristic ATP-binding domains (37), suggesting a role in coupling energy to the transport process. ORF1 to ORF5 also showed homology to the nikABCDE cluster of B. suis (24 to 32% identity) (18) and that of V. parahaemolyticus (33 to 41% identity) (26). They also share 34 to 43% identity with the dipeptide transport system encoded by the dpp operon from Streptococcus pyogenes (27). Furthermore, the Y. pseudotuberculosis gene cluster exhibits the highest homology of all (51 to 72% identity) with the oxd-6 operon from Salmonella enterica serovar Typhimurium: it has been supposed that this operon encodes a nickel or peptide transporter (38), but the matter has yet not been clearly resolved. Finally, the putative ORF1 product was found to be 70% identical to that of an unidentified ORF (also referred to as Orf-1) which follows the regulatory ureR gene present on a 160-kb plasmid in a urease-producing, pathogenic E. coli strain (10). On the whole, these features suggest that the five ORFs code for a nickel transport system, and we thus named them yntA (for Yersinia nickel transport), yntB, yntC, yntD, and yntE, respectively. To prove that they were involved in Ni²⁺ uptake, we used the same approach as for the ureH gene. When plasmid pFS78 bearing the yntABCDE operon was introduced into the E. coli nikA mutant HPX72 harboring the Y. pseudotuberculosis urease cluster borne on plasmid pFS50, ureolytic capacity was restored. In addition, nickel uptake was mediated by the Ynt system at a higher initial rate than that by the UreH system or the E. coli Nik transporter (Fig. 4A).

The Ynt transport system is essential for urease activation in Y. pseudotuberculosis.

To assess the physiological role of the Ynt system in Y. pseudotuberculosis, a ynt-deficient mutant was constructed from wild-type strain 32777, as for the MYUH mutant. Complete deletion of the ynt operon was obtained by using plasmid pFSΔO (Table 1). The genotype of the resulting ynt mutant MYNT was confirmed by PCR assays with the primer sets N9-C1 and N10-C2 and by Southern blot hybridization with appropriate DNA probes (Fig. 2). After growth in LB broth, the ynt mutant's urease activity was dramatically reduced (by 99%) compared with that of the wild-type strain (0.050 ± 0.001 versus 4.8 ± 0.5 μM NH₃/min/mg of protein; Student's t test, P < 0.05) and was completely abolished by concomitant inactivation of ureH (≤0.01 μM NH₃/min/mg of protein). These data indicate that, in contrast to the ureH gene, the ynt genes play a major role for urease activation in Y. pseudotuberculosis.

Both Ynt and UreH are nickel-specific transporters.

To determine the respective contribution of each Y. pseudotuberculosis nickel transporter, the ability of the single or double ureH and ynt mutants to take up nickel was compared with that of the parental strain (Fig. 4B). Accumulation of nickel by the MYOU double mutant was below the threshold of the assay. Uptake in the single mutants MYUH and MYNT can therefore be ascribed to nickel transport by the sole remaining transport system present in these strains. Time course experiments demonstrated that the MYUH mutant's Ynt system gave a significantly (variance analysis: F test, P < 10⁻³) higher rate of uptake than did the MYNT mutant's UreH transporter.

To test the specificity of Ynt and UreH, the effects of cadmium, cobalt, copper, manganese, and zinc ions on nickel uptake by each nickel transporter-deficient Y. pseudotuberculosis mutant were investigated. The competing metal ions were added to a final concentration ∼30-fold greater than that of Ni²⁺ ions (i.e., 5 μM versus 150 nM). None of the metal ions caused significant inhibition of nickel entry into either of the mutants (data not shown). Finally, to estimate the affinity of each transporter for Ni²⁺ ions, transport assays were conducted by using NiCl₂ concentrations ranging from 25 to 150 nM. The K_T value of each transporter was estimated from the initial (linear) uptake rates (<2 min) at 50 ± 5 nM (mean value ± standard deviation of two independent determinations) for UreH and 70 ± 15 nM for Ynt.

DISCUSSION

Nickel transporters play a critical role in the biosynthesis of nickel-dependent enzymes such as hydrogenases and ureases (12). For the first time, we demonstrate the presence of two transporters specific for nickel in a prokaryote. The human and animal pathogen Y. pseudotuberculosis possesses (i) an ABC transporter encoded by the yntABCDE gene cluster, which shows some degree of similarity to the Nik system in several gram-negative bacteria (2, 18, 24, 26), and (ii) a single-component carrier encoded by the ureH gene, displaying homology to polypeptides of the nickel-cobalt transporter family exemplified by HoxN (13). Complete suppression of nickel transport in the ynt ureH double mutant indicated that, apart from Ynt and UreH, there are no other Ni²⁺-specific transport systems in Y. pseudotuberculosis. H. pylori also has two transport systems for nickel acquisition: the high-affinity nickel carrier protein NixA (22) and an ABC transporter (17). However, the latter has not been proven to be specific for nickel.

Database searches revealed that genes homologous to ureH and ynt are present in the genome of the other two pathogenic Yersinia species, Y. pestis (designated by “p”) and Y. enterocolitica (designated by “ent”). Deduced amino acid sequences from ureH_p and ureH_ent are 99 and 93% identical, respectively, to that of the ureH product from Y. pseudotuberculosis, whereas putative Ynt_p and Ynt_ent complexes share 99 and 91.6% of residues, respectively, with Ynt proteins from Y. pseudotuberculosis. Phylogenetic analysis of the putative periplasmic nickel-binding protein encoded by the yntA gene from these three pathogenic Yersinia species showed that YntA, YntA_p, and YntA_ent cluster with the serovar Typhimurium nickel-binding-protein-related Oxd-6a polypeptide and also with the orf1 gene product from urease-producing E. coli strains. Hence, besides the Nik transport system group, the bacterial nickel-ABC transporter family includes another subclass, with Ynt as a prototype carrier and two other members produced by Salmonella and some E. coli isolates.

For several bacterial species in which nickel-transport systems were characterized, genes specifying these carriers were found within or in proximity to genetic loci encoding the nickel-requiring enzyme urease (1, 2, 19, 26) or hydrogenase (13). In addition, nickel permeases of the ABC family are encoded by gene clusters which are harbored either on a plasmid (10), on a pathogenicity island (26), or adjacent to an insertion sequence (2, 26). In Y. pseudotuberculosis, ureH flanks the yut gene (30), located just downstream of the urease accessory gene ureD, whereas the yntABCDE polycistronic unit is upstream of the ureA structural urease gene. The fact that the intergenic space (1,198 bp) between yntE and ureA has a low G+C content (34 versus 47% for the whole Y. pseudotuberculosis genome) and includes many repeated sequences is noteworthy. Furthermore, a copy of insertion sequence IS285 is present 1,848 bp upstream of the Y. pestis yntA gene. Taken together, these genetic features suggest that the chromosomal region containing the ynt operon has been the site of DNA recombination events and that the nickel ABC transport system could have been acquired by yersiniae through horizontal gene transfer.

Nickel uptake assays with Y. pseudotuberculosis ureH and ynt mutants revealed that cellular entry of this divalent cation occurs principally via the Ynt ABC transporter. At the nickel concentration used for the assay, initial rates of nickel uptake by ynt and ureH mutants reached approximately 15 and 60%, respectively, of the wild-type value (Fig. 4B), although the K_T values of UreH and Ynt are similar. This discrepancy could be due to better production of Ynt under our in vitro bacterial growth conditions. Surprisingly, urease activity was not found to correlate with nickel accumulation inside the cells, since it was shown to be strongly reduced (99%) after ynt inactivation but did not significantly differ from that of wild-type Y. pseudotuberculosis after ureH knockout, regardless of nickel or magnesium concentrations in the growth medium. This was not due to a polar effect of the ynt mutation on the downstream ure locus, since urease activity of the ynt mutant was fully restored after trans-complementation with the ynt operon. These discrepancies between the mutants' ureolytic and nickel uptake capacities could be due to regulation by nickel concentration of the Y. pseudotuberculosis urease gene cluster expression, as has been recently demonstrated for H. pylori urease, which is induced by Ni²⁺ at the transcriptional level (35). Mutation of ynt would thus reduce the intracellular nickel concentration below the threshold necessary for induction.

Although weakly homologous (between 25 and 32% identity) to E. coli Nik permease, the Y. pseudotuberculosis Ynt complex is functionally interchangeable with this ABC transporter. However, E. coli cells incorporated much more nickel when expressing ynt instead of the nik gene cluster (Fig. 4A). In the same heterogenous genetic background, UreH is also functional and is as efficient as the endogenous Nik transport system (Fig. 4A). The differences in these systems' nickel transport capacities may reside in their expression in E. coli and could also be linked to their conformation in the cell membrane.

The production of redundant nickel-specific permeases by yersiniae emphasizes the importance of the penetration of this divalent cation into the cell in relation to the biosynthesis of urease—and possibly that of other nickel-dependent enzymes. It also poses the question of their physiological role and raises the possibility that the two systems are expressed under different growth conditions at various stages of the life cycle.