Two Zinc Uptake Systems Contribute to the Full Virulence of Listeria monocytogenes during Growth In Vitro and In Vivo

Abstract

We report here the identification and characterization of two zinc uptake systems, ZurAM and ZinABC, in the intracellular pathogen Listeria monocytogenes. Transcription of both operons was zinc responsive and regulated by the zinc-sensing repressor Zur. Deletion of either zurAM or zinA had no detectable effect on growth in defined media, but a double zurAM zinA mutant was unable to grow in the absence of zinc supplementation. Deletion of zinA had no detectable effect on intracellular growth in HeLa epithelial cells. In contrast, growth of the zurAM mutant was significantly impaired in these cells, indicating the importance of the ZurAM system during intracellular growth. Notably, the deletion of both zinA and zurAM severely attenuated intracellular growth, with the double mutant being defective in actin-based motility and unable to spread from cell to cell. Deletion of either zurAM or zinA had a significant effect on virulence in an oral mouse model, indicating that both zinc uptake systems are important in vivo and establishing the importance of zinc acquisition during infection by L. monocytogenes. The presence of two zinc uptake systems may offer a mechanism by which L. monocytogenes can respond to zinc deficiency within a variety of environments and during different stages of infection, with each system making distinct contributions under different stress conditions.

INTRODUCTION

Competition between a host and pathogen for essential nutrients is a key battleground in infectious disease. The host and pathogen have evolved competing strategies to either limit or acquire the relevant nutrients, and the result of this competition may decide the outcome of infection. This battle for resources is particularly evident with metal ions that act as essential cofactors for numerous essential proteins. Zinc is a good example, being an essential nutrient that serves both structural (e.g., zinc finger) and catalytic (e.g., Lewis acid) cofactor roles. Zinc-binding proteins constitute ca. 6% of the total proteome in Escherichia coli (2), and the cellular zinc quota has been determined to be 0.2 mM, which is the same as that measured for iron and calcium, indicating similar cellular requirements (21). However, despite the high abundance of zinc in the cell, the intracellular availability of free zinc must be strictly controlled (33). Indeed, the femtomolar sensitivity of the E. coli cytoplasmic zinc sensors ZntR and Zur, which regulate zinc efflux and uptake, respectively, implies cytosolic free zinc levels of less than one atom per cell (21).

There is increasing evidence for the importance of zinc during infection. First, it has been demonstrated that there is a rapid decrease in the level of available zinc in plasma in response to bacterial endotoxin and cytokines (11). Second, it has recently been shown that zinc availability regulates the innate immune response to polymicrobial sepsis via NF-κB (3). Third, the cytosolic and lysosomal availability of zinc in phagocytic cells appear to be coordinately regulated in response to infection in order to reduce the availability of zinc to bacteria (17). Pathogenic bacteria have evolved mechanisms to acquire sufficient zinc, or redistribute intracellular zinc, in order to cope with zinc-limited environments that may be generated inside the host as a consequence of infection (13, 15, 19). The contribution of zinc uptake systems to the virulence of several bacterial pathogens has recently been described. For example, the ZnuABC uptake system is required for the adhesion of E. coli O157:H7 to epithelial cells (12) and for the full fitness of uropathogenic E. coli during urinary tract infection (29). Mutants of the intracellular pathogen Salmonella enterica serovar Typhimurium lacking the periplasmic zinc-binding component ZnuA and accessory protein ZinT show impaired growth inside CaCo-2 epithelial cells, with reduced virulence in mice (1, 24). Furthermore, there is elevated expression of ZnuA in S. enterica serovar Typhimurium isolated from both infected macrophage and CaCo-2 cells or from the spleens of infected mice (1). This is consistent with S. enterica serovar Typhimurium experiencing zinc deficiency within the modified phagosome of host cells. The observation that mutants of Yersinia pestis defective for zinc uptake were unaffected in a mouse model of virulence (9) would indicate that the availability of zinc to the pathogen will be determined by the niche occupied within the host.

Listeria monocytogenes is a facultative intracellular bacterium that is also ubiquitous in the environment. This bacterium is the etiological agent of listeriosis, a serious food-borne infection that principally affects the old, immunocompromised, pregnant women and neonates. After uptake or invasion into host cells, L. monocytogenes escapes the phagosome into the cytosol where it is able to nucleate and polymerize host cell actin (30). This enables the bacterium to spread from cell to cell and evade the host innate immune system. The strict regulation of free Zn within the eukaryotic cell means that the cytosolic niche presents its own challenges to L. monocytogenes in terms of acquisition of sufficient zinc for cellular function. Previously, genes encoding a single deduced zinc importing ATPase (zurAM) had been identified in L. monocytogenes within an operon encoding a predicted Zur regulatory protein (8, 22). We demonstrate in the present study a role for ZurAM in zinc uptake and show that its expression is autoregulated by Zur in response to zinc and that it is important for the growth of L. monocytogenes in HeLa cells. We also identify a second related zinc permease system encoded by lmo0153-lmo0154-lmo0155 (herein designated zinABC), which is involved in zinc uptake and is also regulated by Zur. Mutants lacking both systems are unable to grow in defined medium in the absence of zinc supplementation. The observation that neither single mutant was affected for growth in this defined medium indicates at least some functional redundancy between them. Importantly, mutants lacking both zinc uptake systems were severely attenuated for growth in epithelial (HeLa) cells and were defective in actin polymerization and cell-to-cell spread. Loss of either zinc uptake system attenuated virulence in an oral mouse model. Therefore, zinc acquisition is crucial for intracellular replication within the cytoplasm of the infected cell and the survival of L. monocytogenes within animal hosts. The presence of two related ABC-type zinc permease systems may offer a mechanism by which L. monocytogenes can respond to zinc deficiency within a variety of environments and during different stages of infection, each system possibly making distinct contributions under different stress conditions.

MATERIALS AND METHODS

Bacterial strains and culture media.

L. monocytogenes serovar 1/2a strain EGDe InlA^m (34) was used as the wild type. Gene disruption and lacZ insertion mutants generated in the present study are detailed in Fig. 1. The primers used to generate them are shown in Table S1 in the supplemental material. Strains were cultured in either tryptone soy broth (TSB; Oxoid) or a defined medium supplemented with 1% (vol/vol) Casamino Acids as described previously (7) using AnalaR-grade reagents. Glassware was soaked overnight in 4% (vol/vol) nitric acid (Suprapur grade; VWR) before being extensively washed in Milli-Q water. The zinc content of this medium was determined to be 16.87 ± 7.68 μg liter⁻¹ (mean ± the standard error [SE], n = 3) by inductively coupled mass spectrometry (ICP-MS). Glycerol stocks of L. monocytogenes for use in infections were prepared by culturing the bacteria in TSB until absorbance at 600 nm (A₆₀₀) reached ∼0.6, followed by three washes in equal volumes of sterile phosphate-buffered saline (PBS) and resuspension in PBS containing 10% (vol/vol) glycerol.

Fig 1.

Schematic of the zinc uptake operons of L. monocytogenes EGDe (A) and schematic of strains used in the present study (B).

Cloning procedures were performed in E. coli strain DH5α, which was cultured in Luria-Bertani (LB) medium. Where necessary, erythromycin was used at 300 μg ml⁻¹ for E. coli or 5 μg ml⁻¹ for L. monocytogenes. Plasmid pZur was generated in two steps. Initially, 20 pmol of each of two synthetic oligonucleotides matching the sequence of the zur promoter region (Pzur, −108 to −1 relative to the ATG of zurA) were denatured in 10 mM Tris-HCl (pH 8) at 95°C and allowed to anneal by cooling to room temperature at a rate of 1°C min⁻¹ in a thermal cycler. The annealed oligonucleotides were then ligated into SmaI-digested pUNK1 (26) to form plasmid pUNK1-Pzur. The zur gene was then amplified from L. monocytogenes chromosomal DNA by using the primers zurR-F (5′-AACGGTGGATCCTCTTAATTTCAAAACGAGGTGAGG-3′) and zurR-R (5′-AAGCGACTGCAGTTAAGAGGCTTGTAAACATTCTGG-3′). The PCR product was digested with BamHI and PstI (underlined) and ligated into similarly digested pUNK1-Pzur, and the resulting construct was checked by DNA sequencing using primers pUNK1-F (5′-AGAGTTGGTAGCTCTTGATCCGG-3′) and pUNK1-R (5′-TTCAGCAACAATTTTAAACTGC-3′).

Generation of L. monocytogenes mutants, chromosomal complementation strains, and lacZ fusions.

In-frame deletions of zurAM, zinA, and zur were generated by amplifying ∼600 bp of the immediate upstream and downstream regions of the gene(s) to be deleted by PCR using the primers (see the additional information in Table S1 in the supplemental material) and L. monocytogenes chromosomal DNA as a template. The amplified PCR products were digested with the appropriate restriction enzymes (underlined in Table S1 in the supplemental material) and simultaneously ligated into the temperature-sensitive shuttle plasmid pAUL-A (6). Successful clones were confirmed by DNA sequencing by using the primers M13-F (5′-CGCCAGGGTTTTCCCAGTCACGAC-3′) and M13-R (5′-TCACACAGGAAACAGCTATGAC-3′). After transformation into L. monocytogenes (23), integration of the plasmid, and double recombination as described previously (6), only the translational start, stop, and fewer than 15 codons of each open reading frame remained. Successful generation of mutants was confirmed by colony PCR. Nondisruptive transcriptional fusions of lacZ to the zurA-zurM-zur and zinABC operons were generated by inserting a promoterless lacZ between the stop codon of zur or zinC and their respective transcriptional terminators. Regions flanking the insertion site were cloned into pAUL-A and the lacZ gene from plasmid pRS415 (31) was inserted between them, after which standard plasmid integration and double recombination procedures were followed (6). The lacZ gene was amplified incorporating 25 bp of the 5′ untranslated region of L. monocytogenes rpoB, including the Shine-Dalgarno sequence, immediately upstream of lacZ for efficient translation in L. monocytogenes as described previously (7). Insertional inactivation of zur using lacZ (Δzur::lacZ strain [Fig. 1]) was achieved in a similar fashion, except that lacZ was inserted into the BamHI site between the zur flanking regions in plasmid pAUL-AΔzur, followed by transformation into wild-type L. monocytogenes and standard integration and double recombination as described above. β-Galactosidase assays were performed as described previously (7). A single-copy complementation of the ΔzurAM ΔzinA strain (Fig. 1) was constructed by inserting the zurA, zurM, and zinA genes under the control of Pzur at the tRNA^Arg site (18) in three stages. First, regions flanking the tRNA^Arg site were amplified by PCR from chromosomal L. monocytogenes DNA using the primers Xcomp-L1, -R1, -L2, and -R2 (see Table S1 in the supplemental material) and cloned into pAUL-A to form plasmid pAUL-AXcomp. Second, Pzur, zurA, and zurM were amplified with the primers ZurXcomp-F (5′-TGCAGTAGATCTTTGTCTAATTTTTTAATTCG-3′) and ZurXcomp-R (5′-TGCAGTGAATTCTTTTCATGAGTGCTTCAGTTGC-3′), and zinA was amplified with 0153Xcomp-F (5′-TGCAGTGAATTCAAAAAATGCATGGTTTTCTCC-3′) and 0153Xcomp-R (5′-TGCAGTAGATCTGTATTTCATTTGATCGCTTCC-3′). The PCR products were digested (restriction sites underlined) and ligated into the BamHI site of pAUL-AXcomp. The entire complementation construct was then transferred to the chromosome of ΔzurAM ΔzinA as described above and confirmed by colony PCR.

Determination of metal quotients.

Overnight cultures grown in defined medium were diluted 1:100 into fresh defined medium with or without 1 μM N,N,N′,N′-tetrakis(2-pyridylmethyl)ethylenediamine (TPEN; Sigma) and grown at 37°C with shaking at 200 rpm for 22 h. Growth was monitored by measuring the A₆₀₀ and by viable counting. Metal quotas were determined by harvesting 50 ml of 22-h-old cultures and washing them three times with 10 mM Tris-HCl (pH 7.5)–1 mM EDTA and once with Milli-Q water. Cell pellets were dried overnight at 80°C and dissolved in 70% (vol/vol) nitric acid, and the zinc, iron, and manganese contents determined by ICP-MS. Metal content was expressed as atoms per CFU.

Cell cultures and in vitro infection studies.

Eukaryotic cell lines were cultured at 37°C in humidified air with 5% CO₂. HeLa M cells were maintained in Dulbecco modified Eagle medium (DMEM; Sigma) containing 10% fetal bovine serum (FBS; Sigma). Prior to initiating the infection, wells containing 5 × 10⁵ HeLa cells were washed twice with prewarmed PBS before addition of 1 ml of serum-free DMEM containing the appropriate L. monocytogenes strain at a multiplicity of infection (MOI) of 10:1. The zinc content of serum-free DMEM was determined by ICP-MS to be 153.16 μg liter⁻¹. After a 2-h incubation at 37°C, the wells were washed three times with prewarmed PBS, and one set of wells were lysed in 1 ml of cold PBS–0.5% Triton X-100 to obtain a 0-h sample. A 2-ml portion of serum-free DMEM containing gentamicin at 10 μg ml⁻¹ (to kill extracellular bacteria) was added to the remaining wells, and incubation continued at 37°C until the desired time points were reached. Wells were then washed three times with prewarmed PBS, HeLa cells were lysed with cold PBS–0.5% Triton X-100, and the number of viable intracellular L. monocytogenes was determined by serial dilution and viable counting on TSB agar plates.

Plaque assay for cell-to-cell spread.

HeLa M cells (1.5 × 10⁵) were seeded into each well of a six-well plate and incubated for 2 days at 37°C in DMEM containing 10% FBS. The wells were then washed twice with prewarmed PBS, followed by the addition of 1 ml of serum-free DMEM containing a multiplicity of infection (MOI) of 0.1 of the appropriate strain of L. monocytogenes. Incubation was continued at 37°C for 2 h, followed by one wash with 2 ml of prewarmed PBS. Serum-free DMEM (2 ml) containing 1% Sea-Plaque (Lonza) agarose and 10 μg of gentamicin ml⁻¹ was added to each well, allowed to set, and then overlaid with 1 ml of serum-free DMEM. Incubation was continued at 37°C for 3 days. Wells were then stained with neutral red to aid visualization of the plaques (20).

Immunofluorescence assays.

HeLa cells were grown and infected as described above on glass coverslips. At 4.5 or 24 h after infection, coverslips were placed in 3% (wt/vol) paraformaldehyde in PBS for 20 min at room temperature, followed by three washes in PBS. During the second wash, glycine was included at a final concentration of 10 mM. HeLa cells were then permeabilized with PBS containing 0.1% Triton X-100, washed three times with PBS, and then stained for 20 min with fluorescein isothiocyanate (FITC)-labeled phalloidin (Sigma) and Hoechst 33342 DNA dye (Thermo Scientific) in PBS containing 0.5 mg of bovine serum albumin ml⁻¹. Coverslips were washed three times with PBS before application of Mowiol 4-88 (Sigma) and examination under a fluorescence microscope. Images were collected on an Olympus BX51 upright microscope using a 10×/0.30 plan FLN objective and captured using a CoolSnap ES camera (Photometrics). Specific band-pass filter sets for DAPI (4′,6′-diamidino-2-phenylindole), FITC, and Texas Red were used to prevent bleed-through from one channel to the next. Images were then processed and analyzed using ImageJ (http://rsb.info.nih.gov/ij).

Virulence of L. monocytogenes strains in mice.

MF1 outbred mice (Harlan, United Kingdom) were purchased at 7 weeks and housed for a week before infection. Oral gavage was performed as described previously (34), except that bacterial suspensions were inoculated intragastrically in mice that had not been starved, and inoculation was performed using a 21-gauge soft cannula attached to a 1-ml syringe and a dose of 1.5 × 10⁹ CFU ml⁻¹. Mice in a severely lethargic state were culled immediately by cervical dislocation and the time of death recorded. Statistical analysis was performed using the Gehan-Breslow-Wilcoxon test.

RESULTS AND DISCUSSION

Expression of the zurA-zurM-zur operon is zinc responsive.

The zurA and zurM genes are predicted to encode ATPase and membrane permease components of an ABC-type transporter, respectively, while zur encodes a predicted DNA-binding Fur-family transcriptional repressor. Previously, expression of the zurA-zurM-zur operon was shown to be induced by EDTA supplementation in a zinc-dependent manner (8). To further examine the zinc-responsiveness of this operon, a nondisruptive chromosomal fusion of the operon to lacZ was generated (strain zur-lacZ; Fig. 1), and the resulting strain was grown in defined medium in the presence or absence of the zinc chelator TPEN (600 nM), ZnSO₄ (50 μM), or both. The data in Fig. 2 show that there is sufficient zinc in the defined medium to fully repress transcription of the zurA-zurM-zur operon. Supplementation with TPEN resulted in a 4-fold increase in β-galactosidase activity that was abolished upon addition of ZnSO₄ (Fig. 2), which is consistent with transcription of the operon being zinc responsive.

Fig 2.

Transcription of the zurA-zurM-zur operon is regulated by Zur and zinc. The β-galactosidase activity generated by zur-lacZ, Δzur::lacZ, and Δzur::lacZ pZur strains was determined after 22 h of growth in either defined medium alone (black) or in defined medium supplemented with 50 μM ZnSO₄ (dark gray bars), 600 nM TPEN (light gray bars), or both (white bars), as indicated. The data are the means of three independent experiments ± the SE.

To test whether or not the zinc responsiveness of zurA-zurM-zur is due to the action of Zur in L. monocytogenes, a mutant was generated by insertion of lacZ into zur (Δzur::lacZ strain; Fig. 1) with the concomitant removal of 90% of the zur coding region. Elevated β-galactosidase activity was detected in this strain compared to the zur lacZ strain, and expression was not influenced by the addition of ZnSO₄ or TPEN to the growth medium (Fig. 2). Provision of zur in trans on plasmid pZur restored zinc-dependent regulation of the operon (Fig. 2). These data are therefore consistent with L. monocytogenes Zur acting as a zinc-responsive repressor that autoregulates transcription of the zurA-zurM-zur operon, such that transcription is repressed by Zur in the presence of zinc, whereas under conditions of zinc depletion (by TPEN) Zur-mediated repression is relieved.

Elevated expression of zurA-zurM-zur under conditions of zinc deficiency indicates a role for the ZurAM system in zinc import. Hence, strains deleted for zurAM would be expected to accumulate less zinc than wild-type cells when grown in defined medium. To test this hypothesis, an isogenic ΔzurAM strain was generated (Fig. 1), and its total zinc content was compared to that of the wild type following growth in defined medium with or without TPEN supplementation. Surprisingly, no difference was detected between the zinc contents of the two strains (Fig. 3). Manganese and iron levels were also unaffected (Fig. 3). A possible explanation for these findings is that L. monocytogenes possesses an alternative means of acquiring zinc that may compensate for the absence of ZurAM during growth in defined medium.

Fig 3.

Metal content analysis of L. monocytogenes. ICP-MS analysis of nitric acid-digested, chelate-washed cell pellets of L. monocytogenes wild type (WT), ΔzurAM, and ΔzinA strains following growth in defined medium with (+) or without (−) 1 μM TPEN supplementation for 22 h. Values are the numbers of atoms of manganese (black bars), iron (gray bars), or zinc (white bars) per CFU. The data are the means of three independent experiments ± the SE.

Characterization of a second listerial zinc uptake system.

In order to identify other genes with possible roles in zinc uptake, a PatternSearch of the L. monocytogenes genome (http://genolist.pasteur.fr/listilist) was performed using a consensus Gram-positive Zur binding site, [TAAATCGTAATNATTACGATTTA] (22). As predicted from the observed Zur-dependent repression of the zurA-zurM-zur operon (Fig. 2), a Zur binding site was identified 89 bp upstream of the zurA start codon. A second candidate site was identified 39 bp upstream of lmo1882, one of two genes encoding S14 ribosomal subunits in L. monocytogenes (the other being rpsN). Lmo1882 shares 73% sequence identity with the ribosomal protein YhzA in Bacillus subtilis, the expression of which is under the control of Zur. YhzA, like Lmo1882, lacks the two CXXC metal-binding motifs present in its cognate zinc-requiring protein RspN and enables continued ribosome assembly by functionally replacing YhzA during zinc limitation (13, 19). A further Zur binding site was identified upstream of the lmo0153-lmo0154-lmo0155 operon. The products of this operon bear similarity to the AdcA lipoprotein (57% identity), AdcC ATPase (57% identity), and AdcB membrane permease (50% identity) components, respectively, of the Streptococcus pneumoniae AdcABC high-affinity zinc uptake system (4, 10, 16, 22). Lmo0153 contains a His/Asp/Glu-rich region characteristic of zinc transporters (15) and also contains the three conserved zinc-coordinating His residues that have been identified in ZnuA of E. coli and Synechocystis and the B. subtilis metal-binding lipoprotein YcdH (15). We therefore hypothesized that the lmo0153-lmo0154-lmo0155 operon encodes a second zinc uptake system in L. monocytogenes, and we propose renaming the operon zinABC.

To test whether or not the zinABC operon is part of the Zur regulon and responsive to zinc, a single-copy, nondisruptive transcriptional fusion of the operon to lacZ was generated in both a wild-type and a Δzur background (zin-lacZ, Δzur zin-lacZ; Fig. 1), and the β-galactosidase activity generated by both strains was assessed monitoring the growth in defined medium (Fig. 4). As with zurA-zurM-zur, basal expression of the zinABC operon was found to be very low, but expression was strongly upregulated following growth in the presence of TPEN, an effect that could be reversed by cosupplementation with ZnSO₄ (Fig. 4). Furthermore, deletion of zur (Δzur zin-lacZ) resulted in constitutively high β-galactosidase expression under all conditions, and normal regulation was restored by providing zur in trans (Fig. 4). These data therefore confirm that Zur represses zinABC expression in the presence of zinc. To examine the role of zinABC in zinc uptake, ICP-MS analysis of chelate-washed extracts of ΔzinA mutant again showed that, following growth in the presence or absence of TPEN, there was no difference in the zinc content of the mutant compared to the wild-type strain (Fig. 3). Crucially, in contrast to the single ΔzurAM or ΔzinA mutants, which showed normal growth in defined medium (Fig. 5A), the double ΔzurAM ΔzinA mutant lacking both systems was unable to grow in defined medium (Fig. 5A). The failure of the double mutant to grow in defined medium prevented the metal content of this strain being determined. The growth defect of the double mutant could be complemented by providing the zurAM and zinA genes in trans in a single copy (ΔΔcomp, Fig. 1), under the control of Pzur, or by supplementation of the medium with 50 μM ZnSO₄ (Fig. 5B and C).

Fig 4.

Transcription of the zinABC operon is regulated by Zur and zinc. The β-galactosidase activity generated by zin-lacZ, Δzur zin-lacZ, and Δzur zin-lacZ pZur strains was determined after 22 h of growth in either defined medium (black bars) or in defined medium supplemented with 50 μM ZnSO₄ (dark gray bars), 600 nM TPEN (light gray bars), or both (white bars), as indicated. The data are the means of three independent experiments ± the SE.

Fig 5.

Growth of zinc uptake mutants in vitro. Growth (A₆₀₀) of wild-type (squares), ΔzurAM (triangles), ΔzinA (diamonds), or ΔzurAM ΔzinA (circles) strains in defined medium (A) or in defined medium supplemented with 50 μM ZnSO₄ (B) or TSB (D). Final A₆₀₀ values of cultures of wild-type (WT), ΔzurAM ΔzinA mutant, or the complemented ΔzurAM ΔzinA mutant (ΔΔcomp) strains were determined in defined medium after 22 h of growth (C) or TSB after 5 h of growth (E) either with (white bars) or without (black bars) 50 μM ZnSO₄ supplementation. In all cases, the data are the means of at least three independent experiments ± the SE.

TSB, which contains approximately 10 μM zinc, was able to support growth of the ΔzurAM ΔzinA mutant, although with an increased doubling time compared to the wild type (62 and 46 min, respectively; Fig. 5D). Complementation of the double mutant or additional supplementation of the medium with 50 μM ZnSO₄ restored wild-type-like growth to the double mutant (Fig. 5E). Taken together, these data show that L. monocytogenes possesses two zinc uptake systems that are upregulated under conditions of zinc deficiency and form part of the Zur regulon. Either zurAM or zinABC is required for growth in defined medium in the absence of zinc supplementation. Under conditions of increased zinc availability in the growth medium (≥10 μM), L. monocytogenes is apparently able to acquire sufficient zinc to support growth even in the absence of both uptake systems.

ZurAM is the principal zinc uptake system during intracellular growth of L. monocytogenes.

A previous in vivo transcriptomic study had revealed that zinA is upregulated during L. monocytogenes infection of mice (5). To assess the contribution of the ZurAM and ZinABC zinc uptake systems to the intracellular growth of L. monocytogenes, HeLa cells were infected with the wild-type or the ΔzurAM, ΔzinA, or double ΔzurAMΔzinA mutant strains, and growth was monitored at intervals over the course of 24 h. A Δhly ΔplcB mutant, which is unable to escape the phagosome following invasion, was included for comparison (14). Equivalent numbers of wild type and ΔzinA mutant were recovered from the infected HeLa cells up to 24 h postinfection (Fig. 6). However, the number of ΔzurAM mutant recovered was significantly reduced compared to the wild type (24 h, P < 0.001, n = 4), indicating that growth of this mutant is impaired in HeLa cells. The number of the double ΔzurAM ΔzinA mutant remained unchanged from 1 h to 24 h postinfection and was statistically indistinguishable from the number of Δhly ΔplcB mutant recovered (Fig. 6). Taken together, these data indicate that the double ΔzurAM ΔzinA mutant was incapable of growth in the cytoplasm of HeLa cells.

Fig 6.

Comparison of intracellular growth of zinc uptake mutant strains in HeLa cells. HeLa cells were infected with an MOI of 10 with either the wild-type (squares), ΔzurAM (triangles), ΔzinA (diamonds), ΔzurAM ΔzinA (circles) or Δhly ΔplcB (crosses) strain, and growth was assessed by lysing HeLa cells at intervals, followed by serial dilution and viable counting of bacteria. The data are the means of at least three independent experiments ± the SE.

The ΔzurAM ΔzinA strain is defective in actin polymerization and cell-to-cell spread.

To determine whether the failure of the double zinc uptake mutant to grow in HeLa cells was a consequence of it being trapped in a vacuole within the cell, immunofluorescence microscopy was used to monitor actin accumulation and polymerization by the wild-type, ΔzurAM ΔzinA, and Δhly ΔplcB strains during infection (Fig. 7). At 4.5 h postinfection, many wild-type cells were found either coated in actin or at the head of actin comet tails, indicating vacuolar escape and actin-based intracellular motility (Fig. 7C and D). As anticipated, no actin accumulation by the Δhly ΔplcB strain could be observed at 4.5 h postinfection (Fig. 7E and F), indicating that this strain remained trapped in the primary vacuole following invasion. Similar numbers of the ΔzurAM ΔzinA mutant compared to the wild type were found to be coated in actin; however, relatively few comet tails were observed for the former (Fig. 7G and H). Indeed, a more detailed investigation of an average of five fields in each of three independent experiments revealed the number of ΔzurAM ΔzinA bacteria associated with comet tails to be 8-fold lower than observed for the wild type (Table 1). These data indicate that the double zinc uptake mutant is not defective in escaping the primary vacuole or in nucleating actin, but in transitioning from actin nucleation to actin polymerization. To ensure that the onset of actin tail formation was not simply delayed in the double mutant, these experiments were repeated 24 h after infection (Fig. 7I to K). However, while there was an abundance of wild-type L. monocytogenes associated with actin comet tails, no detectable increase in the number of actin-tail forming ΔzurAM ΔzinA bacteria was observed (Fig. 7K).

Fig 7.

Immunofluorescence analysis of HeLa cells at 4.5 h (A to H) or 24 h (I to K) postinfection. Images are of uninfected HeLa cells (A, B, and I) or HeLa cells infected with a wild-type (C, D, and J), Δhly ΔplcB (E and F), or ΔzurAM ΔzinA (G, H, and K) strain. Bacteria appear red due to labeling of the bacterial cell surface by the anti-listeria antiserum, cellular F-actin appears green due to phalloidin-FITC, and HeLa cell nuclear material appears blue due to staining with Hoechst 33342 DNA dye.

Table 1.

Intracellular phenotypes of wild-type and ΔzurAM ΔzinA mutant strains

Parameter at 4.5 h postinfection	Straina
	Wild type	ΔzurAM ΔzinA mutant
No. of fields counted	5 ± 1	5 ± 1
Listeria/HeLa cell ratio	16.1 ± 1.8	4.6 ± 0.4
Actin-coated bacteria (% total bacteria)	24.42 ± 0.97	25.25 ± 3.2
Comet tails (% total bacteria)	2.16 ± 0.24	0.26 ± 0.29
Comet tails (% actin-coated bacteria)	9.05	1.16

^aValues are presented as means ± the standard error, where applicable (n = 3).

A likely consequence of defective actin tail formation is a failure of the bacterium to spread from cell to cell, as has been observed in mutants with altered actA expression (35), or loss of polar ActA accumulation, as observed in a cwhA mutant (25). Indeed, the ΔzurAM ΔzinA mutant was unable to form plaques in a homologous plaque assay (Table 2). To our knowledge, a direct role for zinc in ActA-mediated actin polymerization has never been reported. However, an inability to acquire zinc is likely to have pleiotrophic effects, and several examples of filamentous forms of the ΔzurAM ΔzinA mutant were apparent during immunofluorescence microscopy at 4.5 and 24 h (Fig. 7H and K). This could indicate that the defect in actin tail formation is related to impaired cell division, and hence ActA polarization, as previously observed (25). It is also possible that the low level of intracellular replication of the ΔzurAM ΔzinA mutant is contributing to the lack of actin polymerization. An alternative explanation for the failure of cell-to-cell spread could be reduced activity of the phosphatidylcholine phosphoesterase, PlcB. PlcB is required for escape of L. monocytogenes from the double-membrane vacuole that results from intercellular spread (32). However, PlcB activation is dependent upon the zinc metalloprotease, Mpl (27). It is therefore conceivable that an inability to acquire sufficient zinc might limit the amount of mature PlcB produced by L. monocytogenes, thus impeding cell-to-cell spread and subsequent growth of the bacterium.

Table 2.

The ΔzurAM ΔzinA mutant is defective in intracellular spread

Strain	No. of plaquesa
Wild type	139 ± 26.7
ΔzurAM ΔzinA mutant	0
Δhly ΔplcB mutant	0

^aValues are presented as means ± the standard error (n = 3).

Both zinc uptake systems are required for full virulence in a mouse model of infection.

To test the importance of each zinc uptake system during infection by L. monocytogenes, mice were infected by oral gavage and their survival postinfection recorded. Deletion of zurAM caused a significant reduction in virulence with substantially increased survival of mice infected with the mutant compared to the wild type (P = 0.006, n = 10; Fig. 8). Deletion of zinA also significantly attenuated the virulence of L. monocytogenes (P = 0.05, n = 10; Fig. 8). The double mutant was also attenuated and significantly less virulent than the wild type (P = 0.04, n = 10; Fig. 8) but not significantly different to the zinA mutant (P = 0.87, n = 10). No significant difference in survival was detected between mice infected with the single uptake mutants. These data demonstrate that both zinc uptake systems are required for L. monocytogenes virulence in the oral mouse model. The observation that the double mutant was least attenuated appears curious. One possible explanation is that in contrast to either single mutant, the double mutant induces comparable levels of interleukin-1β (IL-1β) release from primary macrophages to that seen with the wild type (unpublished observations). As a consequence, it is possible that the double mutant induces a significantly greater inflammatory response than that of either single mutant, and this increase in inflammation may explain the reduced attenuation of the double mutant. The reasons for the observed IL-1β release are not yet clear but are part of our ongoing studies.

Fig 8.

Survival of MF-1 mice orally infected with either wild-type L. monocytogenes (solid line) or ΔzurAM (dashes), ΔzinA (dots and dashes), or ΔzurAM ΔzinA (dots) mutant strains. Mice were infected as described in Materials and Methods and monitored over a 240-h period.

Concluding remarks.

Our results demonstrate that two ABC permease systems, ZurAM and ZinABC, are involved in zinc uptake in L. monocytogenes. Mutants lacking both systems are unable to grow in defined medium in the absence of zinc supplementation. Both systems are expressed in response to zinc limitation and form part of the Zur regulon in this bacterium. We show that both systems are required for the full virulence of L. monocytogenes in HeLa cells and a mouse model of infection. Intriguingly, there is some functional redundancy between the two systems that varies depending upon the environment in which the bacterium is growing. For example, deletion of either system reduces virulence in mice, whereas the ZurAM system appears to play a greater role during growth in the cytoplasm of HeLa cells. These studies formally demonstrate the importance of zinc uptake by L. monocytogenes for virulence, and the presence of two specific uptake systems may allow access to different sources of zinc within different environments. To our knowledge, the presence of two specific zinc uptake systems is rare in bacteria. Recently, a second zinc uptake system (ZevAB) was identified in nontypeable Haemophilus influenzae in addition to the ZnuABC system (28). In this case the Zev system was shown to be important for growth in the lungs (28). The important role of zinc in the onset of the host innate response would suggest that the battle for zinc availability between the pathogen and the host during the early stages of an infection will have implications in deciding the outcome of the pathogen-host interaction.