A Putative P-Type ATPase Required for Virulence and Resistance to Haem Toxicity in Listeria monocytogenes

Regulation of iron homeostasis in many pathogens is principally mediated by the ferric uptake regulator, Fur. Since acquisition of iron from the host is essential for the intracellular pathogen Listeria monocytogenes, we predicted the existence of Fur-regulated systems that support infection. We examined the contribution of nine Fur-regulated loci to the pathogenicity of L. monocytogenes in a murine model of infection. While mutating the majority of the genes failed to affect virulence, three mutants exhibited a significantly compromised virulence potential. Most striking was the role of the membrane protein we designate FrvA (Fur regulated virulence factor A; encoded by frvA [lmo0641]), which is absolutely required for the systemic phase of infection in mice and also for virulence in an alternative infection model, the Wax Moth Galleria mellonella. Further analysis of the ΔfrvA mutant revealed poor growth in iron deficient media and inhibition of growth by micromolar concentrations of haem or haemoglobin, a phenotype which may contribute to the attenuated growth of this mutant during infection. Uptake studies indicated that the ΔfrvA mutant is unaffected in the uptake of ferric citrate but demonstrates a significant increase in uptake of haem and haemin. The data suggest a potential role for FrvA as a haem exporter that functions, at least in part, to protect the cell against the potential toxicity of free haem.


Introduction
Iron is indispensable for the growth of most bacteria, serving as a cofactor for enzymes involved in essential metabolic pathways such as glycolysis, DNA synthesis, energy generation, and detoxification of oxygen radicals [1,2]. The correlation between iron acquisition and bacterial virulence has been well documented [3,4,5] and the absolute requirement for this metal for both host metabolism and bacterial growth results in significant competition for iron in the host [6]. Following bacterial infection host responses are evoked which sequester iron, making it relatively unavailable for bacterial metabolism [7].
In the Gram positive intracellular pathogen Listeria monocytogenes, iron deficient environments have been shown to upregulate the expression of the principal virulence regulator PrfA and significantly increase the production of the haemolysin Listeriolysin O promoting phagosomal escape, and the actin polymerisation protein ActA which plays a role in cell-to-cell spread [8,9,10]. It has been hypothesized that the requirement for iron has played a part in driving the evolution of an intracellular life-cycle for L. monocytogenes as the bacterium can utilize the iron-saturated protein ferritin stored in the cytosol of host cells (as reviewed by McLaughlin et al. [11]).
As iron-limiting conditions can be encountered in both the natural environment and during host infection, free-living pathogenic bacteria such as L. monocytogenes have evolved mechanisms to acquire iron from a variety of sources. Iron acquisition is mediated by a number of distinct systems that have been characterized in L. monocytogenes: a citrate inducible receptor for the uptake of ferric citrate, utilization of exogenous siderophores, catechol siderophore-like molecules, and catecholamine complexes, and iron acquisition via a cell-surface transferrinbinding protein [12]. A comprehensive analysis of the iron acquisition systems in L. monocytogenes identified a variety of iron sources which can be used for growth, including eukaryotic ironbinding proteins (haemoglobin, ferritin, transferrin and lactofer-rin), ferric siderophores (enterobactin and corynebactin) and iron complexes of hydroxymates (ferrichrome, ferrichrome A, and ferrioxamine B) [2]. In addition, the same study also identified two genetic loci responsible for the uptake of ferric hydroxymates and haemin/haemoglobin. Deletions in fhuD or fhuC strongly reduced ferrichrome uptake and a deletion in hupC eliminated uptake of haemin and haemoglobin and resulted in decreased virulence potential [2]. However, it is clear that many other loci putatively involved in iron homeostasis in L. monocytogenes remain to be characterized by functional genetics approaches [13,14].
Maintaining a balanced acquisition of iron from the external environment is essential for bacterial survival. Whilst pathogens must compete for iron during infection excess intracellular iron can lead to the generation of toxic hydroxyl radicals via the Fenton reaction. Iron homeostasis in most bacteria, including L. monocytogenes, is controlled by the regulatory protein Fur (ferric uptake regulator) or a functional equivalent [15]. In the presence of sufficient levels of iron, Fur acts as a repressor whereby an iron-Fur complex prevents gene transcription by binding to specific Fur-box sequences upstream of the start codon of target genes [16].
Recently Ledala et al. [17] used DNA microarray analysis to examine gene expression in a Fur mutant and identified Furregulated genes in L. monocytogenes, including genes encoding iron transporters and proteins involved in iron storage. In this study, we undertook an independent genome-wide search to identify putative Fur-box consensus sequences in the genome of L. monocytogenes. This approach identified a number of candidate Fur-regulated loci, including some (such as lmo0641) that were not identified previously through microarray analysis [17]. We undertook a systematic functional analysis of selected Furregulated loci by creating plasmid-insertion mutants and subsequently testing these for virulence potential in the murine model. This led to the identification of Fur-regulated virulence factor A, FrvA (encoded by frvA/lmo0641), a novel Fur regulated virulence factor which is absolutely required for growth of L. monocytogenes under restricted iron conditions and for systemic infection. We carried out iron uptake studies on the frvA mutant and determined that the mutant demonstrates a significant increase in uptake of haem and is also sensitive to elevated haem concentrations. Sensitivity to haem toxicity may account for the significant attenuation of virulence during the systemic phase of infection in the murine infection model.

Results and Discussion
In silico identification of putative Fur regulated genes Fur has been identified as a major regulator of iron homeostasis in numerous Gram-positive and Gram-negative bacteria [16,18,19]. Regulation of iron uptake is particularly important during infection as pathogens must scavenge iron from sources in the host organism. Indeed, deregulation of iron uptake through elimination of Fur has been shown to significantly impact upon virulence potential in a number of pathogenic bacteria, including L. monocytogenes [20,21]. Surprisingly, recent approaches to identify novel in vivo-induced genes in L. monocytogenes (such as microarray and IVET approaches) have failed to identify the key inducible systems for iron-uptake during infection [22,23,24]. In addition, signature tagged mutagenesis approaches have also failed to identify the mechanisms of intracellular iron uptake in this pathogen [25]. We therefore employed a systematic functional genetic analysis of selected Fur-regulated genes and identified a locus (lmo0641, now designated frvA) that is absolutely required for the systemic phase of L. monocytogenes infection.
Ledala and coworkers have recently utilised microarray analysis to identify members of the Fur regulon in L. monocytogenes [17]. We concurrently used the classical 19 bp Fur-binding motif defined in B. subtilis [26] ( Figure 1A) to mine the L. monocytogenes EGDe genome for similar motif sequences. We used two primary criteria to limit the number of sequences identified. Firstly, the identified sequence should be within 350 bp of an annotated start codon and secondly, a match at 16 or more of the 19 positions was required. Anything less than 16/19 was not considered unless the annotated ORF was deemed to have a likely role in iron acquisition based on bioinformatic analysis. This approach identified a subset of the Fur-regulated loci determined through microarray analysis [17]. However, we also identified Fur-regulated loci at lmo2431 (previously identified as a potential Fur-regulated locus by Jin et al. [2] and lmo0641 (the subject of this study) which were not detected using the cut-off criteria employed by Ledala et al. [17]. Another locus (lmo0484) was identified here which is adjacent to a gene (lmo0485) identified using microarrays and therefore may form part of an operon.
The loci identified as containing Fur-binding motifs are represented in Figure 1. In each case, where the Fur box was upstream of a putative operon, RT-PCR confirmed co-transcription of all the genes in the operon (data not shown). The Fur boxes were aligned and a graphical display of the results was generated using 'sequence logo' which generates a consensus for Listeria that is identical to that in Bacillus ( Figure 1B) [27]. Fur regulation was confirmed through RT-PCR analysis of representative genes in both wild-type L. monocytogenes and a Dfur mutant. The results validated the microarray data described previously [17] and also confirmed that lmo2431 and lmo0641 are regulated by Fur.

Virulence analysis of plasmid insertion mutants
We created mutants using the pORI19 integration strategy as this method is relatively rapid, results in stable mutations and lends itself to analysis of a large number of loci in a reasonable timeframe [20,28]. Two of the identified loci (lmo1007 and lmo0484) consisted of a single small (,500 nt) gene and were considered too small for plasmid disruption and were not analysed here. Mutation of fri has been described elsewhere [29,30,31]. Where the Fur box was upstream of an operon we chose the first open-reading frame for plasmid disruption as this would increase the likelihood of causing pleiotropic effects on co-transcribed downstream genes. Plasmid disruptions at the correct locations were confirmed by PCR, using a primer based on the EDGe chromosome and one based on the plasmid. The absence of the repA gene in mutant strains selects against excision and extrachromosomal maintenance of the integration plasmid, ensuring stable integrants for subsequent analysis (see experimental procedures for details). mRNA was extracted from each of the mutants and RT-PCR analysis confirmed that plasmid disruption of the target gene was associated with the complete elimination of expression from each locus with the exception of the lmo2431 mutant (a locus previously analysed by Jin et al. [2]) in which gene expression was greatly reduced (data not shown).
In this initial screen, three of the mutants in Fur-regulated loci exhibited a significant reduction in virulence potential relative to the wild type (P,0.05) ( Figure 1E). The most significantly affected mutant in this screen was pORI19::frvA.

The Fur-regulated virulence (frvA) locus is required for effective infection
To confirm an essential role for frvA in the virulence of L. monocytogenes two precise in-frame deletion mutants were created (see experimental procedures). An initial mutant was created Recent studies have suggested that a more accurate representation of the Fur box is that of a 7-1-7 motif. The 19 bp sequence was used to search the Listeria monocytogenes EGDe genome sequence (Listilist). (B) Identified sequences were aligned and a graphical display of the results was generated using the web based programme sequence logo (17). (C) Genetic organisation of 29 putative Fur regulated genes (black/gray) at 12 chromosomal loci. All genes are drawn approximately to scale using the L. monocytogenes EGDe genome sequence data. Lmo numbers refer to the National Centre for Biotechnology Information annotation scheme. Fur boxes are represented by black circles. Gray genes indicate those disrupted in EGDe in the course of this study. Lollipops are used to illustrate putative stem loop terminator regions. (D) RT-PCR analysis was used to confirm Fur through the deletion of the central region of the lmo0641 gene, from residues 85-416 inclusive (mutant designated DfrvA  ). As toxicity has previously been associated with the generation of truncated membrane proteins through partial deletion [31] we also created a precise deletion mutant in which the entire open reading frame was deleted. This mutant was designated DfrvA. Both mutants were complemented using the pPL2 plasmid to reintroduce a single copy of frvA (designated DfrvA::pPL2frvA and DfrvA  ::pPL2frvA). Although growth of DfrvA was unaffected in nutrient-rich media (BHI), this mutant was recovered at significantly lower levels (three-log reduction) from the spleens of infected mice on day three post-infection when compared to the wild-type. Numbers recovered from the liver at three days post inoculation indicated a lesser, but still significant reduction in levels as compared to the wild-type. The reintroduction of frvA fully restored the virulence potential ( Figure 2A). These data definitively establish a critical role for frvA in L. monocytogenes virulence potential and pathogenesis. Notably, DfrvA  ) also demonstrated a dramatic reduction in virulence potential in the murine model ( Figure S1).
Larvae of the Wax Moth (Galleria mellonella) have recently been utilized as an alternative pathogenicity model for L. monocytogenes [32,33]. Here we also analysed the virulence potential of DfrvA using this model system ( Figure 2B). While no deaths were observed over time for the insects that received PBS, a significant difference was seen between insects receiving the wild-type and the DfrvA mutant. The LT-50 for insects infected with the wild-type strain was 26 hours while over 50% of the DfrvA-infected insects were still alive after 43 hours. The significance of iron acquisition to the virulence of bacterial pathogens has previously been investigated in this insect model. Work by Daou et al. [34] demonstrates a role for IlsA, a surface protein in Bacillus cereus that binds haemoglobin and ferritin, to pathogenesis in the G. mellonella host. In order to determine the possible influence of the downstream gene lmo0642 in murine virulence of L. monocytogenes we created an in-frame mutation in this locus. Interestingly, this gene is apparently not required for pathogenesis. Deletion of lmo0642 failed to affect the ability of L. monocytogenes to reach either the liver or spleen in mice infected by the intraperitoneal (i.p.) route ( Figure 2C) or to grow intracellularly (data not shown).
DfrvA was compared to the wild-type and complemented strains for their ability to replicate within J774 macrophage cells ( Figure  S2). After one hour DfrvA displayed no significant difference in invasion of J774 cells when compared to the wild-type or complement strains. Subsequent readings taken after three, five and seven hours represent intracellular survival and multiplication of these strains within the cell line. Similarly, no significant difference was observed in the ability of DfrvA to survive inside J774 macrophages over time, as all three strains displayed growth of approximately one log after 7 hours.

Bioinformatic analysis of frvA
FrvA is a putative transmembrane protein consisting of six predicted transmembrane regions (SOSUI) and is annotated as being similar to a heavy metal-transporting ATPase (http:// genolist.pasteur.fr/ListiList/). The closest non-listerial homologues reside in Bacillus spp. A predicted heavy metal-transporting ATPase in B. megaterium was found to share 56% identity and 72% similarity (over a query coverage of 621/626 amino acids) with FrvA. A predicted cadmium-transporting ATPase in B. halodurans C-125 also shared close homology with 55% identity and 72% positives (over 618/626 amino acids) (NCBI Blast). Three conserved domains were identified in FrvA using the Conserved Domain Search from NCBI including a P-type ATPase superfamily, a haloacid dehalogenase-like (HAD) hydrolase, and a cation transport ATPase. In addition, FrvA was found to contain several classic P-type ATPase motifs such as the phosphorylation motif D 321 KTGTLT and the hinge region motif G 518 DGIND. Similar to other type I heavy metal-transporting ATPases such as YkvW in Bacillus subtilis and CtpA, a P-type ATPase involved in copper homeostasis in L. monocytogens, FrvA also possesses both an M4 motif S 277 PC and an HP motif, S 358 LHPLA, respectively [35].
Lmo0642, the product of the downstream gene, is also predicted to be localized to the bacterial membrane (PSORT) and also has 6 transmembrane regions (SOSUI). No conserved domains were identified (NCBI) and its closest homolog is a hypothetical protein (EF0716) from Enterococcus faecalis V583 (NCBI Blast).

Regulation of frvA by Fur
qRT-PCR analysis of the wild-type L. monocytogenes EGDe strain and a Dfur mutant confirmed that lmo0641 is under the negative regulation of Fur. Using the 2 2DDCt method to calculate the relative changes in gene expression, lmo0641 was shown to be upregulated 93-fold in Dfur compared to the wild-type. Transcription of frvA was also found to under the control of PerR, a Fur homolog which functions as an Fe(II)-dependent peroxide stress sensor and which regulates putative metal transport and storage functions [36]. In addition to the classical Fur box a putative PerR binding region was identified upstream of the annotated start codon of frvA. De-repression of frvA was seen in the absence of either regulator. However, no further impact was observed in a DfurDperR double mutant (data not shown). The significance of this dual regulation by Fur and PerR is unclear, but highlights some degree of interplay between these two regulators. It is interesting to note that frvA (lmo0641) was also previously determined to be regulated by PrfA, the master regulator of virulence gene expression in L. monocytogenes [37]. The locus is not preceded by a detectable PrfA binding motif but the authors noted the presence of a binding site recognized by Sigma B, the general stress response regulator. Taken together, the evidence suggests that the locus is a member of multiple regulatory networks, perhaps reflecting the importance of FrvA in L. monocytogenes niche adaptation.
DfrvA displays increased haemin uptake and elevated sensitivity to haem toxicity In an attempt to understand the virulence defect displayed by DfrvA we carried out extensive physiological analysis of the mutant strain. A [ 59 Fe]-citrate uptake assay indicated that the ability of DfrvA to acquire ferric citrate was not impaired when compared to the wild-type L. monocytogenes strain ( Figure 3A). Both strains transport ferric citrate with similar affinity (K m ) and velocity regulation of all identified genes and to give an indication of the increase in expression levels. Control primers were used to ensure that template cDNAs were of equal concentration. Samples were removed at various cycles of PCR (cycle number in brackets) and visualised on agarose gels. A repeat experiment demonstrated similar results. Results were also verified through real-time PCR analysis. (E) In vivo survival of disruption mutants in Fur-regulated loci in the murine infection model. Mice were injected i.p. with either the wild-type or mutants and the number of bacteria recovered from the spleen was determined three days post-inoculation. Error bars represent the standard deviations from the mean (n = 4). * indicates means are significantly different to the wild-type (P,0.05). ND, not detected. doi:10.1371/journal.pone.0030928.g001 (V max ). Although Adams et al. [38] have reported that a citrate inducible iron uptake system exists in L. monocytogenes we demonstrate here that the FrvA system is not involved in the direct uptake of ferric citrate. The existence of an iron reductase has previously been suggested in L. monocytogenes based upon physiological data [39,40] although this remains the subject of some debate [2]. We performed iron reductase assays but could find no significant difference between wild-type and mutant cells in ability to reduce iron in these assays, suggesting that this locus does not encode an iron reductase (see Table S1).
During infection free iron is not available to bacterial cells whereas haem (Hb) and haemin (Hn) represent a potentially abundant source of iron [41]. However haem can be relatively toxic to cells at elevated concentrations [42,43]. We investigated the rates of haemin uptake in DfrvA and observed significant differences between the wild-type and mutant strains in the acquisition of [ 59 Fe]-Hn ( Figure 3B). The rate of haemin transport by DfrvA (V ma x = 30.6 pMol per 1610 9 cells per minute) was nearly twice that of the wild-type strain (Vmax = 18.8 pMol per 1610 9 cells per min). Subsequent analysis of the mutant in iron-limiting MOPS-L media supplemented with haemoglobin and haemin revealed that L. monocytogenes DfrvA displayed growth behavior distinct from that of the wild-type and complement strains ( Figure 4). Growth of the wild-type and complement was restored upon addition of increasing concentrations of Hb and Hn (0.2 and 2.0 mM) to iron-limiting media ( Figures 4A, 4C, 4F). In contrast, growth of DfrvA required addition of 0.2 mM Hn and Hb, whereas a higher concentration of 2.0 mM was shown to reduce growth suggestive of toxicity ( Figures 4B and 4E). Nutrition tests were performed to assess the capability of the strains to utilize iron from several different sources. DfrvA displayed no impairment in ability to utilize ferric siderophores, Hb or Hn when compared to the wild-type and complement strains ( Figure 5).
As FrvA displays homology to bacterial heavy-metal transporting ATPases and with the knowledge that cation-transporting ATPases function in maintaining cation homeostasis [35], we investigated the sensitivity of DfrvA to toxic levels of heavy metal sulfates such as copper, cobalt, cadmium, and zinc as well as iron. Exposure to a disk that contained 1 M FeSO 4 resulted in a larger zone of clearance in DfrvA when compared to the wild-type, indicative of elevated toxicity. However sensitivity to other heavy metals such as CdSO 4 , CoSO 4 , CuSO 4 and ZnSO 4 was comparable in both the wild-type and mutant ( Figure S3). The data suggest that deletion of frvA does not affect the sensitivity of cells to heavy metals such as cadmium, cobalt, copper and zinc but confirms the contribution of this locus to iron homeostasis.

Global disruption of iron homeostasis in the DfrvA mutant
As physiological analysis of DfrvA revealed iron-related phenotypes, we investigated the possibility that deletion of this locus could lead to altered expression of other genes in the L. monocytogenes genome involved in iron homeostasis. qRT-PCR was used to evaluate the differential expression of three ironrelated genes in the wild-type and mutant strains ( Figure 6). We chose two Fur-regulated genes; lmo2186 which encodes a homologue of SaulsdC and bears homology to a haemin binding protein IsdC in S. aureus [44], and lmo1959, designated as fhuD encoding the L. monocytogenes ferrichrome binding protein [2,44]. In addition, lmo2431 (hupD) was also analyzed as this gene is part of the hupDGC operon encoding an ABC transporter involved in uptake of haemin and haemoglobin [2,44]. qRT-PCR analysis revealed a strong induction of both lmo2431 and lmo1959 in DfrvA compared to the wild-type strain. lmo2431 was shown to be up- regulated 210-fold and lmo1959 up-regulated 164-fold in the mutant strain. lmo2186 also displayed a an induction in DfrvA, with an almost 5-fold difference observed between the wild-type and mutant. As Fur is generally considered a repressor of transcription [16], the induction of two Fur-regulated genes in DfrvA is supported by our finding that the fur gene was shown to be down-regulated almost 6-fold in DfrvA.

Conclusions
Using a functional genetics approach we identified a novel Furregulated locus (frvA) in L. monocytogenes that is essential for virulence and for resistance to haem and haemin-mediated toxicity. It is known that L. monocytogenes has the capacity to utilise iron-loaded haemoglobin and haemin as sources of iron [31]. Furthermore, elimination of haemoglobin and haemin uptake through mutation of the HupC transport system significantly impairs virulence potential, indicating that iron acquisition from haem is essential for pathogenesis [2]. However, haem and haemin are known to be toxic for bacteria and many bacteria express specific mechanisms for detoxification of haem [43].
FrvA possesses P-type ATPase and hydrolase conserved domains and is homologous to other heavy-metal transporting ATPases in Staphylococcus and Bacillus. Previous work by Francis and Thomas [35] identified another P-type ATPase, encoded by ctpA, which is involved in copper homeostasis in L. monocytogenes. Significantly, mutagenesis of the ctpA locus resulted in a strain that was unaffected in intracellular growth in the J774 macrophage cell line, but was impaired in ability to cause infection in the murine model [45]. Although P-type ATPases are known to mediate the transport of various heavy metals in bacteria, iron transport is most often associated with the structurally unique ATPases of the ABC transporter family [46]. However, Mta72, a P-type ATPase in M. tuberculosis, has been shown to transport iron transferred from the siderophore carboxymycobactin and is another rare example of a P-type ATPase involved in iron homeostasis [6].
It is interesting to note that the HrtA system in S. aureus also functions as a haem exporter and deletion of hrtA in that organism causes dysregulation of Fur expression resulting in pleiotrophic effects [47]. Whilst HrtA is an ABC transporter rather than a Ptype ATPase we note homologies between FrvA and HrtA (21% identity over 221 amino acids). Certainly the results presented here suggest functional similarities between FrvA and HrtA though further experimental work will be necessary to directly compare both systems.
We did not demonstrate a role for FrvA in transport of ferric citrate or in iron reduction by L. monocytogenes and the mutant was not impaired in intracellular growth in vitro. Rather the predominant phenotype of DfrvA is an increased uptake of haemin and significantly increased sensitivity to both haemin and haemoglobin toxicity and reduced virulence during systemic infection. However we acknowledge that further work is necessary to determine the precise biochemical mechanisms underpinning FrvA activity. The profound dysregulation of iron homeostasis in DfrvA results in the de-repression of other Fur-regulated loci which complicates interpretation of the analysis of the mutant and which may necessitate the future use of isolated liposomal protein models to delineate its precise function.

Ethics statement
All animal procedures were approved by the University Research Ethics Board (UREB) in University College Cork (approval ID 2008/32) and were carried out in a specialized , and in BHI broth (F). The bacteria were cultured in BHI broth overnight. In A-E they were then subcultured in MOPS-L to stationary phase, and at t = 0 subcultured again at 1% into MOPS-L containing different concentrations of Hb or Hn. In F, at t = 0 they were subcultured into BHI broth. The flasks were shaken at 37uC and absorbance at 600 nm (initially close to zero for all cultures) was monitored for 12-26 h (note different scales). Because of the slow growth of L. monocytogenes in iron-restricted minimal media, this graphic representation focuses on the comparison of the mutant strains at later times in the growth cycle. doi:10.1371/journal.pone.0030928.g004 facility. Work was carried out under license from the Irish Department of Health.

Bacterial strains, plasmids and culture conditions
Listeria monocytogenes strains were grown in Brain Heart Infusion (BHI) (Oxoid) broth at 37uC and Escherichia coli strains were grown in Luria-Bertani (LB) broth at 37uC. Strains and plasmids used in this study are listed in Table 1. For solid media, agar (1.5%) was added. Antibiotics, obtained from Sigma Chemical Company, were added in the following concentrations; 50 mg/ml ampicillin for pKSV7 in E. coli and 10 mg/ml chloramphenicol for pKSV7 in L. monocytogenes.

DNA manipulations
Gel extraction was performed using the Qiagen Gel Extraction Kit (Qiagen). Plasmid DNA isolation was carried out utilizing the Qiagen QIAprep Spin Miniprep Kit (Qiagen). PCR reagents and T4 DNA ligase, supplied by Roche Diagnostics GmbH (Mannheim, Germany), and restriction enzymes (New England Biolabs) were all used according to the manufacturer's instructions. Oligonucleotide primers were synthesized by MWG and are listed in Table S2. PCR reactions were completed using a PTC-200 (MJ Research) PCR system. Colony PCR was performed following lysis of cells with IGPAL CA-630 (Sigma). Genomic DNA was isolated from L. monocytogenes using a chromosomal kit (Sigma) according to the manufacturer's instructions.

Creation of plasmid insertion mutants
A central portion of the gene of interest was amplified by PCR and cloned into the multiple cloning site of pORI19 (RepA 2 ) [20,28]. Following plasmid isolation, electrotransformation of L. monocytogenes EGDe containing pVE6007 (RepA + /Temperature sensitive) was performed according to the protocols outlined by Park and Stewart, (1990). Loss of pVE6007 was achieved by transferring 10 ml of a 30uC overnight culture to BHI broth prewarmed to 42uC with subsequent growth for 16 hrs at 42uC and isolation on prewarmed BHI-Em (5 mg/ml) agar plates at 42uC. Loss of pVE6007 (Cm s ) was confirmed by replica plating onto BHI-Em (5 mg/ml) and BHI-Cm (10 mg/ml) plates with overnight incubation at 30uC. Integration results in the formation of a stable Em r mutant and was confirmed by PCR using a primer  outside the region of integration and a primer specific to the plasmid.

Construction of deletion mutants
As described by Horton et al. [48] the Splicing by Overlap Extension (SOE) procedure was utilized to create a complete gene deletion mutant. This is an in-frame, non-polar deletion of a gene in the L. monocytogenes EGDe chromosome. Two pairs of primers were designed, SOEA/SOEB and SOEC/SOED, to amplify two fragments of approximately equal size on either side of the gene to be deleted using the proofreading enzyme Vent polymerase (New England Biolabs). These AB and CD products were then gel extracted to ensure purity, mixed in a 1:1 ratio, and were spliced together using SOEA/SOED primers. This AD product was digested and cloned into pKSV7, a temperature sensitive plasmid. The resulting construct was electroporated into competent E. coli DH5a cells and transformants were selected on Luria-Berani plates with ampicillin. The plasmid was isolated using the Qiagen QIAprep Spin Miniprep kit. The presence of the correct insert was verified by sequencing (Lark Technologies Inc., Essex, UK) using the pKSV7 MCS primers M13F and M13Rmut. The isolated plasmid was electroporated into competent L. monocytogenes EGDe cells. Transformant selection took place on Brain-Heart Infusion agar containing chloramphenicol. Clones in which chromosomal integration of the plasmid had occurred are selected by serial passaging at 42uC and are streaked onto pre-warmed BHI-Cm agar plates. Continuous passaging at 30uC in BHI broth followed by replica plating onto BHI and BHI-Cm plates ensures plasmid excision. Chloramphenicol sensitive colonies were tested for gene deletion using primers SOEX, located upstream, and SOEY, located downstream of the gene of interest.

Complementation of deletion mutants
A site-specific phage integration vector, pPL2, was used for the complementation of SOEing deletion mutants. This vector integrates within the tRNA Arg gene on the chromosome. Vent polymerase (New England Biolabs), a proof reading enzyme, was used to amplify the entire deleted gene and flanking regions, including the upstream gene promoter. Primers CompA and CompB included restriction sites corresponding to those on the MCS site of pPL2. The PCR product was gel extracted to ensure purity, and was digested and cloned into pPL2. The resulting construct was electroporated into competent E. coli DH5a cells and transformants were selected on Luria-Berani plates with chloramphenicol. The plasmid was isolated using the Qiagen QIAprep Spin Miniprep kit. The presence of the correct insert was verified by sequencing (Lark Technologies Inc., Essex, UK) using the pPL2 MCS primers T3F and T7R. The isolated plasmid was electroporated into competent SOEing mutant cells. Transformant selection took place on Brain-Heart Infusion agar containing chloramphenicol. The presence of the gene was authenticated using a forward running check primer that anneals to the middle of the gene and SOE D, located on the cloned insert. Integration of pPL2 to the correct site was confirmed using primers PL102, located upstream of the integration site, and the SOE D primer.

RNA extraction
Total RNA was extracted using both the Macaloid Clay method, outlined by Raya et al. [49], and the Qiagen RNeasy Mini Kit. Cultures were grown overnight shaking at 37uC. A 1% inoculum was added to 30 mLs BHI broth and cultures were grown at 37uC until an OD 600 nm of 0.3 was reached. 30 mLs of culture were pelleted by centrifugation at 4,000 g for 7 minutes. The supernatant was removed, the pellet was washed with 1 mL cold TE buffer (10 mM Tris, 1 mM EDTA: pH 8.0), and centrifuged again for 13,000 g for 1 minute. Again the supernatant was removed, and the pellet was resuspended in 20 mL lysozyme (50 mg/mL), 400 mL cold TE buffer, and left at room temperature for 3 minutes. Subsequently, the cell suspension was added to a 1.5 mL screw-cap plastic tube containing 50 mL 10% sodium dodecyl sulphate, 500 mL phenol-chloroform (5:1), 175 mL Macaloid Clay and 0.5 g 425-600 mm glass beads (Sigma). Cell disruption was achieved using a bead beater (Mini-beadbeater 8TM cell disrupter, Biospec products.) Cells were beaten for 1 minute, placed on ice for 1 minute, beaten again for 1 minute, and then centrifuged for at 13,000 g for 3 minutes. The organic layer was removed and precipitated with 1:10 volume sodium acetate, and 2.5 volume 96% ethanol at 280uC for 20 minutes. Following this step, samples were put through the Qiagen RNeasy Mini Kit and then eluted in 50 mL TE buffer. RNA samples were treated with RNase-free DNase I set (Qiagen) and DNA-free (Ambion) was used to remove any DNA. The concentration of RNA was quantified utilizing a Nano-Drop (ND-1000 spectrophotometer). A PCR, carried out with 16S rRNA primers; L142 and U141, was used to ensure the absence of DNA in the samples. The reverse transcriptase PCR was run using 4 mL random primer p(dN) 6 , 2 mL RNA, and 2 mL DEPC water (Sigma) at 65uC for 10 min, and put directly on ice. To these samples, 32 mL of a mastermix was added containing; 1 mL Expand Reverse Transcriptase (Roche), 8 mL 56 Buffer (Roche), 4 mL 100 mM dTT (Roche), 1 mL dNTP mix (dATP, dCTP, dGTP, dTTP; 10 mM) and 18 mL DEPC water. This reaction was carried out at 30uC for 10 min, 42uC for 3 hours, and held at 4uC. cDNA was confirmed through PCR using L142 and U141 primers and the wild-type L. monocytogenes extracted DNA as a positive control.

Quantitative real-time PCR
The Universal Probe Library Assay Design Center (https:// www.roche-applied-science.com/sis/rtpcr/upl/adc.jsp) was used to design PCR primers which correspond to a specific probe in the library. Primer sequences and corresponding probes are listed in Table S2. The 16S rRNA gene was used as a housekeeping gene to compensate for any variability in the initial amount of starting total RNA. Amplification reactions consisted of 2.5 mL of cDNA, 6.4 mL of 26 FastStart TaqMan Probe Master (Roche), primers (900 nM) and probe mix (250 nM). RNase-free water was added to bring the total volume of the reaction to 10 mL. Reactions were performed in duplicate on 384-well plates using the LightCycler 480 System (Roche). Negative control reactions, without cDNA, were also included on the plate. Thermal cycling conditions were carried out according to manufacturer's instructions (Roche) and the 2 2DDCt method [50] was used to calculate the relative changes in gene expression from the qRT-PCR experiments.

Growth curves
Growth of Listeria monocytogenes in MOPS-L media. EGD-e and its mutant derivatives were grown in BHI overnight, and then subcultured at 1% into BHI broth or MOPS-L media. In the latter case, the bacteria were grown to stationary phase, and for growth rate determinations they were subcultured again at 1% into MOPS-L containing Hn or Hb at varying concentrations. The cultures were shaken at 37uC and OD 600 nm was monitored at indicated time points up to 26 hours.

Metal disk assay
Cultures were grown overnight shaking at 37uC. A 2% inoculum was added to 10 mL of fresh BHI and cultures were grown to logarithmic phase (0.3OD) at 37uC. 400 mL of log phase cell cultures were added to 4 mL of cooled, molten soft agar (0.75%) and poured on top of a petri dish containing 20 mL BHI agar. After solidifying, a sterilized 13 mm disk (Whatman) was placed on top of the overlay. Metals used were made up in 1 M stocks in which 35 mL of each metal were dispensed onto the center of the disk. The plate was then incubated overnight at 37uC, and the zone of clearance surrounding the disk was measured.

Fe binding and uptake experiments
For binding and transport determinations, we prepared 59 Fe complexes of citrate (specific activity 150 to 1,000 cpm/pMol) and haemin [44]. For 59 Fe-citrate, we provided the organic ligand in 50-fold molar excess. We conducted adsorption and transport experiments [2,44] over a range of concentrations, by adding appropriate amounts of 59 Fe complexes to two aliquots of 2610 7 cells of EGD-e or its mutants, and incubating the aliquots for 15 s and 75 s, respectively, before collecting and washing the cells on 0.2 micron filters. The 15 s aliquot measured the amount initially bound to the cells, which when subtracted from the second timepoint, gave the amount transported during a 1 min period. At each concentration, data were collected in triplicate and averaged. The K d and capacity of 59 Fe-siderophore binding were determined by using the ''Bound-versus-Total'' equation of Grafit 5.09 (Erithacus, Ltd., Middlesex, UK), and K m and V max of transport were calculated using the ''Enzyme Kinetics'' equation.

Macrophage assay
This intracellular survival assay was carried out using J774 mouse macrophage cells (originally obtained from the American Type Culture Collection, Manassas, VA). 24-well tissue culture plates were seeded with 1610 5 live cells per well in DMEM (Gibco) containing 10% fetal calf serum and incubated in 5% CO 2 at 37uC for 40 hours. For infection, bacteria were prepared by centrifuging 1 mL of an overnight culture which was then washed once in PBS, and resuspended in 1 mL DMEM. Bacteria were diluted in DMEM and 1610 7 CFU was added to each well containing washed macrophage cells. To increase contact between macrophages and bacteria, the 24-well plates were centrifuged at 1500 rpm for 10 min and incubated for 1 hour in 5% CO 2 at 37uC. To kill extracellular bacteria, 1 mL of 100 mg/mL gentamycin (Sigma) was added to each well and incubated for an additional 30 min. Bacteria surviving intracellularly were enumerated at time points taken after addition of gentamycin. Each well was washed twice in PBS, cells were lysed with 250 mL ice cold water containing 0.02% Triton X (Sigma), and scraped in a similar manner using a pipette tip. Serial dilutions were carried out on the lysate and plated on BHI agar overnight at 37uC.

Murine virulence assay
Cultures were grown overnight with shaking at 37uC. Cultures were centrifuged, washed in PBS (Sigma), resuspended and diluted to 1610 6 CFU/mL in PBS. BALB/c mice were inoculated with 4610 5 CFU in 200 mL PBS intraperitoneally (i.p.). The mice were euthanized 3 days post-infection. Spleens and Livers were harvested and then homogenized in PBS. Bacteria were enumerated by plating the serial dilutions of organ homogenates on BHI agar left to incubate overnight at 37uC.

Galleria mellonella virulence assay
Cultures were grown overnight with shaking at 37uC. Cultures were centrifuged, washed, and resuspended in an equal volume of PBS (Sigma). Infection of Galleria mellonella was performed according to the protocol outlined by Joyce et al. [32]. Briefly, insects were obtained from Livefood, UK and were stored in the dark at room temperature prior to use. 3 groups, containing 10 insects per group, were injected with 1610 6 CFU/10 mL of the wild-type L. monocytogenes EGD-e strain (group 1), 1610 6 CFU/ 10 mL EGDeD0641 (group 2), or10 mL PBS (group 3) to serve as a negative control. Bacterial suspensions were injected using a sterile Hamilton syringe and a 30-Gauge disposable needle into the first right pro-leg of the second set of pro-legs. All ten insects per group were placed together in a Petri-dish lined with Whatman paper and incubated in the dark at 37uC. Insects were examined over several days and time of death was recorded.  Figure S3 Metal toxicity disk assay. 35 mL of 1 M copper, cobalt, cadmium and zinc sulfates (A) or iron sulfate (B) were added to a 13 mm disk placed on an overlay of wild-type or D0641 cells grown up to 0.3 OD. Plates were incubated for 24 over night and zones of clearance (ZOC) were measured (mm). No statistical differences were observed between strains in the sensitivity to heavy metals CdSO4, CoSO4, CuSO4 and ZnSO4. While no ZOC was observed around disks containing 1 M FeSO4 for the wild-type, a ZOC of 7.5 mm60.5 mm was seen for DfrvA. Experiments were done in triplicate.