EfaR Is a Major Regulator of Enterococcus faecalis Manganese Transporters and Influences Processes Involved in Host Colonization and Infection

Abstract

Metal ions, in particular manganese, are important modulators of bacterial pathogenicity. However, little is known about the role of manganese-dependent proteins in the nosocomial pathogen Enterococcus faecalis, a major cause of bacterial endocarditis. The present study demonstrates that the DtxR/MntR family metalloregulator EfaR of E. faecalis controls the expression of several of its regulon members in a manganese-dependent way. We also show that efaR inactivation impairs the ability of E. faecalis to form biofilms, to survive inside macrophages, and to tolerate oxidative stress. Our results reveal that EfaR is an important modulator of E. faecalis virulence and link manganese homeostasis to enterococcal pathogenicity.

INTRODUCTION

Enterococci are nosocomial opportunistic pathogens that can cause infections of the urinary tract, bloodstream, intra-abdominal and pelvic regions, and surgical sites (1, 2). They are of concern in the hospital environment because of their ability to bind to, colonize, and produce biofilms on medical devices such as stents, catheters, artificial cardiac pacemakers, and prosthetic heart valves (3). Primarily, these bacteria are human commensals that have adapted to complex environments rich in nutrients and low in oxygen, such as the gastrointestinal tract and the oral cavity. For enterococci to act as pathogens, they must start by adhering to and then invading host tissues. During the process of tissue invasion, enterococci encounter an environment that differs greatly from the site of colonization, with higher redox potentials, limited essential nutrients, phagocytic leukocytes and other host defenses. A potentially limiting factor for bacterial growth is the scarcity of various metals which are also required by the host and which are needed for various essential cellular processes. Metal homeostasis has important implications for enzymatic function and for appropriate transcriptional control of regulatory networks governing gene expression under diverse environmental conditions (4). In particular, acquisition of manganese plays an important role in pathogenesis in a number of bacterial species (5). This metal ion plays a role as a cofactor of enzymes involved in metabolism, in signal transduction, and in protection against oxidative stress (5, 6). Mn²⁺ is present at about 36 μM in saliva (7) but at only nanomolar concentrations at internal sites of the host (8, 9). Manganese could be an important cue by which the new environment is sensed by invading bacteria, functioning as a signal of the transition to internal body sites (10). Similarly, it has been proposed that iron depletion may serve as a signal allowing many bacterial pathogens to sense that they are within a vertebrate host (11).

Little is known about how E. faecalis achieves manganese homeostasis and about the impact of this ion on E. faecalis biology. In the only study relating manganese to E. faecalis pathogenicity, the efaCBA operon of E. faecalis JH2-2 was described as being dependent on EfaR, a regulator belonging to the DtxR/MntR family of transcriptional regulators (12). Binding of EfaR to the promoter region of efaCBA was shown to be promoted by manganese, which acted as a corepressor. The efaCBA operon encodes the adhesion lipoprotein EfaA, first isolated from serum of a patient and described as an endocarditis-associated antigen (13) and later demonstrated to be expressed during enterococcal endocarditis (14). Its role as a virulence factor was established when infection with an E. faecalis OG1RFΔefaA mutant led to delayed mortality in mice (15). efaCB encodes an ATP-binding cassette (ABC) transporter (12).

Previously, we reported on the transcriptomic response of E. faecalis V583 to manganese, zinc, and copper stresses (16). Here, we analyzed the promoter regions of genes that were differentially expressed as a consequence of excess metal added to the medium in which E. faecalis V583 was grown. A DNA motif, here called the EfaR binding motif (EBM), was identified in the promoter regions of 30 genes, five of which were previously shown to be repressed by manganese and copper and induced by zinc stress. Among this group of genes was the efaCBA operon. Thus, we further investigated the role of EfaR as a regulator of genes preceded by the EBM and the involvement of EfaR, and some of the genes it regulates, in a number of processes important for virulence, namely, oxidative stress tolerance, biofilm formation, and survival inside macrophages.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

Bacterial strains and plasmids used in this study are listed in Table 1. E. faecalis strains were grown as standing cultures at 37°C in M17 medium with 0.5% glucose (GM17) or in metal-depleted (Chelex-treated) GM17. Chelated medium was prepared by autoclaving M17 medium with 2% Chelex 100, sodium form (Sigma-Aldrich Chemie GmbH, Steinheim, Germany), followed by 2 h of stirring. Glucose (0.5%) and 20 μM MgSO₄ were added to allow E. faecalis growth in this medium (chelGM17). Escherichia coli was grown in LB (Luria-Bertani) broth or 2× YT (yeast extract and tryptone) broth in a shaking incubator at 37°C. Strains with thermosensitive plasmids were grown at 28°C. Chloramphenicol was used at a concentration of 30 μg/ml for E. faecalis VE14412; tetracycline was used at concentrations of 15 μg/ml for E. coli VE14192 and 10 μg/ml for E. faecalis strains. Kanamycin, for E. coli strains, and erythromycin, for E. faecalis strains, were used at 50 μg/ml. When necessary, 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) (VWR International Ltd., Leicestershire, United Kingdom) was added to the growth medium at 250 μg/ml. BHI (Oxoid, Hampshire, England) plates were used to count viable bacterial cells in the intramacrophage survival and the oxidative stress assays.

Table 1.

Strains and plasmids used in this study

Strain or plasmid	Descriptiona	Reference or source
Strains
Enterococcus faecalis
VE14089	Wild type; V583 cured of pTEF1, pTEF2, and pTEF3	19
SAVE20	Eryr; VE14089 (pSAVE10, carrying Pef0575::lacZ fusion)	This work
SAVE21	Eryr; VE14089 (pSAVE11, carrying PmntH2::lacZ fusion)	This work
VE14412	Cmr; VE14089 (pGhost 3)	19
SAVE22	Tetr; VE14089 efaR mutant with tet insertion in efaR	This work
SAVE23	Eryr; Tetr; SAVE22 (pSAVE10, carrying Pef0575::lacZ fusion)	This work
SAVE24	Eryr; Tetr; SAVE22 (pSAVE11, carrying PmntH2::lacZ fusion)	This work
EF0577_S1_CA	Tetr; VE14089 ef0577 mutant with tet insertion in ef0577	19
EF0579_S1_CA	Tetr; VE14089 ef0579 mutant with tet insertion in ef0579	19
SAVE25	Tetr; Eryr EF0579_S1_CA (pSAVE10, carrying Pef0575::lacZ fusion)	This work
SAVE26	Tetr; Ery; EF0579_S1_CA (pSAVE11, carrying PmntH2::lacZ fusion)	This work
EF2076_S1_JO	Tetr; VE14089 efaA mutant with tet insertion in efaA	19
Escherichia coli
VE14188	Kanr; GM1674 (repA+)	19
VE14192	Kanr; Tetr; GM1674 (pVE14218)	19
Plasmids
pILORI4	Eryr; pIL252 with MCS and promoterless lacZ of pORI13	20
pSAVE10	Eryr; pILORI4 carrying Pef0575::lacZ fusion	This work
pSAVE11	Eryr; pILORI4 carrying PmntH2::lacZ fusion	This work
pG+host 3	Cmr; repA+; thermosensitive replication	43
pVE14218	Tetr; derived from p3TETTery and pOrinew	19
pSAVE12	Tetr; pVE14218 with ca. 80% of efaR in MCS	This work

^aKan^r, kanamycin resistance; Ery^r, erythromycin resistance; Cm^r, chloramphenicol resistance; Tet^r, tetracycline resistance; MCS, multiple cloning site.

DISCLOSE analysis.

DISCLOSE software (17) was used on DNA microarray data obtained previously (16). Genes differentially expressed in the presence of high concentrations of zinc, manganese, and copper ions (16) were examined in order to determine clusters of these genes that exhibit similar expression patterns and identify overrepresented DNA binding sites in the upstream DNA sequences of genes from those clusters.

Construction of an E. faecalis efaR mutant.

Single-crossover insertion mutagenesis was performed to create an E. faecalis efaR mutant using the two-vector system essentially described by Law and coworkers (18). Plasmid pG⁺host 3 and the integrative plasmid pVE14218 were used in this strategy (19). Primers used for construction of the integration vector are presented in Table 2. The efaR mutant was confirmed by Southern hybridization.

Table 2.

Primers used in this study

Primer	Nucleotide sequence (5′–3′)
EF1005-1	TACCTACCTGCAGGAGGGGACTGCTGCTTTAAAGCTGACGG
EF1005-2	TACTACCCTAGGGTGGGCAGTTCTGGACTCGATGTTTCGG
OEF102	GGCGATCGGACTAAACAATTGAACACGGC
OEF106	TGATGAAACGGCACGGATAG
PEF0575fw	GAGAAGAATTCAATGGTCTTCCCATGTATTTAGG
PEF0575rev	AAGCAGGATCCATTTTCCAGCACCATTTGGACC
PmntH2fw	ATTTCGAATTCCTTTAAGACCGCACATTTACG
PmntH2rev	GCATAGGATCCCAAATGATGTCTTTGCTTTGG
pILORI4-1	CCATTCGCCATTCAGGCT
pILORI4-2	CCGCTACGGATCACATCT

Construction of transcriptional lacZ fusions.

Transcriptional fusions with the E. coli lacZ gene were constructed in plasmid pILORI4 (20). Primer pairs PEF0575fw-PEF0575rev and PmntH2fw-PmntH2rev (Table 2) were used to generate PCR fragments spanning the upstream regions of ef0575 and mntH2, respectively. The fragments were cloned into the EcoRI-BamHI sites of pILORI4, yielding pSAVE10 and pSAVE11, respectively. Each plasmid was introduced into E. faecalis VE14089 and its efaR and ef0759 mutants. Confirmation of these constructs was done by PCR with primers pILORI4-1 and pILORI4-2 (Table 2) and subsequent nucleotide sequencing.

β-Galactosidase assays.

Cells were grown in GM17 or chelGM17 with 50 μg/ml erythromycin and metal ions. Metal solutions were used in the following added metal concentrations: ZnCl₂, 0, 0.5, 2, 4, and 6 mM; MnCl₂, 0, 0.2, 0.4, 0.6, and 1 mM; CuSO₄, 0, 0.025, and 0.05 mM, CoCl₂, 0, 0.1, 0.5, and 1 mM; FeCl₂, 0, 0.1, 0.5, and 1 mM; NiSO₄, 0, 0.1, 0.5, and 1 mM; and MgCl₂, 0, 0.1, and 0.5 mM. Cells from the mid-logarithmic growth phase were spun down and frozen in liquid nitrogen to be used later. Frozen cell pellets were resuspended in the same volume of Z buffer (60 mM Na₂HPO₄, 40 mM NaH₂PO₄, 10 mM KCl, 1 mM MgSO₄, and 50 mM β-mercaptoethanol, pH 7.0), and the optical density at 600 nm (OD₆₀₀) was measured. Cells in diluted suspensions were permeabilized with 100 μl of chloroform and 50 μl of 0.1% sodium dodecyl sulfate (SDS). Reactions were started with the addition of 200 μl of 2-nitrophenyl β-d-galactopyranoside (Sigma-Aldrich Chemie GmbH) (4 mg/ml in 0.1 M phosphate buffer, pH 7.0) and stopped with 500 μl of 1 M Na₂CO₃. The reaction mixtures were centrifuged for 5 min at 14,000 rpm, and the ODs at 420 and 550 nm were recorded. The activity of LacZ (in Miller units) was calculated as described by Miller (21). For each assay, three independent replicates were performed. Results were compared using a two-tailed unpaired t test at the 97% confidence interval. P values of <0.03 were considered statistically significant.

DNA microarray experiments and data analysis.

DNA microarray experiments were performed essentially as described previously (16). RNA was isolated from 30 ml of E. faecalis VE14089 and efaR mutant cultures grown to mid-exponential phase in GM17, following the procedure previously described (16). cDNA synthesis and indirect Cy-3/Cy-5-dCTP labeling and hybridization procedures were performed as reported elsewhere (16). Glass slides carrying 70-mer oligonucleotides for 3,160 genes of E. faecalis V583 (22) were scanned in a GeneTac LS IV confocal laser scanner (Genomics Solutions, Huntingdon, United Kingdom). DNA microarray data were obtained from three independent biological replicates hybridized to glass slides with gene oligonucleotides spotted in triplicates.

Slide images were analyzed using GenePix Pro 6.0 software (Axon Instruments, Union City, CA). Data processing was performed as described elsewhere (16). Expression ratios were calculated from the measurements of at least seven of nine spots. Differential expression tests were performed on expression ratios with a local copy of the Cyber-T implementation of a variant of the t test (23).

Intramacrophage survival assay.

The macrophage survival assay was basically performed as described previously (24), with some modifications. Confluent monolayers of the murine macrophage cell line J774.A1, established from a tumor that arose in a female BALB/c mouse, were infected with cultures of E. faecalis VE14089 or the efaR, ef0577, ef0579, or efaA mutant. Approximately 4 × 10⁶ bacteria were added to J774.A1 monolayers, to yield a multiplicity of infection of approximately 10 bacteria per cell. The cell cultures were incubated at 37°C in a 5% CO₂ atmosphere for 1 h to allow bacterial adherence and entry. After this, 250 μg/ml gentamicin and 60 μg/ml penicillin G were added to the cultures, followed by an incubation of 1 h to kill extracellular bacteria. 1% Triton X-100 (Fluka Analytical, Sigma-Aldrich, Buchs, Switzerland) at 1% in phosphate-buffered saline, pH 7.4 (Gibco, Invitrogen, Paisley, United Kingdom) was used to lyse macrophage cells at 0, 2, 4, 6, 8, and 24 h postinfection. Lysates were then diluted and plated on BHI plates to count viable intracellular bacteria. Assays were performed in three independent replicates, and results are reported as the intracellular survival index (SI), i.e., the percent (mean) of the internalized CFU at 0 h postinfection that survived after phagocytosis. Results were compared using a two-tailed unpaired t test at the 97% confidence interval. P values of <0.03 were considered statistically significant.

Oxidative stress assay.

The oxidative stress assay was performed on E. faecalis VE14089 and the efaR, ef0577, and efaA mutants grown in BHI until an OD₆₀₀ of 0.5 was reached. Then, the cells in the cultures were exposed to hydrogen peroxide stress (20 mM H₂O₂ in 0.9% NaCl) as described by Verneuil and coworkers (25). Each data point is the average of the data from three independent experiments with triplicate plating. The percentage of survival at a given time point was calculated by determining the ratio of the number of CFU at a given time after treatment to the number at time zero, both obtained by plate counting using BHI plates. Results were compared using a two-tailed unpaired t test at the 97% confidence interval. P values of <0.03 were considered statistically significant.

Biofilm formation assay.

E. faecalis VE14089 and the efaR, ef0577, ef0579, and efaA mutants were grown as standing cultures for 16 h in 2× YT medium with 0.5% glucose, at 37°C. The cultures were subsequently diluted 1:100 (vol/vol) in the same fresh medium. A 200-μl portion of the diluted cell suspensions was inoculated in sterile 96-well polystyrene microtiter plates (Sarstedt AG & Co., Nümbrecht, Germany). Biofilm formation on polystyrene was quantified after 24 h of incubation at 37°C by the crystal violet-staining method, as previously described (26). For each assay, three independent experiments were performed in six replicates. All experiments included blank wells (medium without inoculum). Results were compared using a two-tailed unpaired t test at the 97% confidence interval. P values of <0.03 were considered statistically significant.

Microarray data accession number.

Microarray data were uploaded to GEO under accession number GSE33698.

RESULTS

Genes carrying an EBM site respond to metals.

Analysis with DISCLOSE (17) of our previous data on the E. faecalis V583 transcriptomic response to high metal concentrations (16) revealed the presence of a putative DNA binding motif in the upstream regions of several genes whose expression was affected by manganese, zinc, or copper addition (Fig. 1). In a search with a cutoff of 10⁻⁴ in the intergenic regions of the V583 genome sequence, more genes were found to carry the same putative metal-responsive motif in their upstream sequences, making a total of 30 genes (see Table S1 in the supplemental material). We labeled this motif the EfaR binding motif (EBM), for reasons explained below. Among others, the EBM is present in the promoter region of the efaCBA operon. In a previous study of this operon, a bioinformatic approach was used to search the genome of E. faecalis V583 for an earlier-described DtxR consensus binding box (12, 27). This allowed the identification of a 14-bp palindromic sequence (TTAGGNNNNCCTAA) in the promoter regions of 13 genes (12). The sequence was found twice in the promoter region of the efaCBA operon; they were labeled box 1 (TTAGGTGCGCCTAA), located upstream of the −35 region, and box 2 (TAAGGCAAACCTAA), immediately upstream of the RBS. Of the 13 genes identified by Low et al. (12), only the efaCBA promoter region carried both palindromic sequences; the promoters of all other genes contained only one. The function of the two boxes in the regulation of the efaCBA operon still needs to be clarified. Our analysis of metal-responsive genes allowed the identification of one motif encompassing box 1, in the promoter region of the efaCBA operon and upstream of the ef0575, ef0579, mntH1, and mntH2 genes. These genes were also identified by Low and coworkers (12). We labeled our empirically identified motif the EfaR binding motif (EBM) because of the previous demonstration of binding of EfaR to a 131-bp DNA fragment encompassing the efaCBA promoter and box 1 (12) and in light of the DNA microarray and β-galactosidase data presented below. The efaCBA operon was previously observed to respond to manganese (12), and we have reported that this operon and four of the other the EBM-containing genes, namely, the ones in Table 3, respond to manganese, zinc, and copper (16). To examine the role of the EBM site in sensing metal stress, we measured the activities of plasmid-encoded lacZ fusions to the promoter regions of mntH2 (pSAVE11; Fig. 2A) and ef0575 (pSAVE10; data not shown, as this plasmid behaves like PmntH2::lacZ), two genes carrying the EBM but never before assessed as to their roles in sensing and responding to metals. Expression of lacZ was significantly affected only upon addition to the medium containing Zn²⁺, Mn²⁺, Cu²⁺, and Fe²⁺ [strain VE14089(pSAVE11)]. Other metal ions, namely, Mg²⁺, Co²⁺, and Ni²⁺, had no significant effects (data not shown). The results confirm the upregulation in the presence of Zn²⁺ and downregulation in the presence of Mn²⁺ and Cu²⁺ of the mntH2 and ef0575 promoter activities observed in the DNA microarray experiments (16). Of these promoters, only the activity of the mntH2 promoter was affected by Fe²⁺, but in a way opposite to that of manganese, as iron induced transcription from PmntH2.

Fig 1.

The EBM, present in the promoter regions of several genes differentially expressed in the presence of excess Zn²⁺, Mn²⁺, or Cu²⁺ ions. The WebLogo is based on the MEME weight matrix and shows the bit score of A, C, T, and G nucleotides at each position of the motif (44).

Table 3.

E. faecalis V583 genes carrying an EBM in their promoter regions that were upregulated in the presence of excess zinc (Zn) and downregulated in the presence of manganese (Mn) or copper (Cu)^a

Geneb	Locus tag	Motif sequence	Descriptionc
ef0575	EF0575	TAGGCTTGACTAAA	Cationic ABC transporter, ATP-binding protein
ef0576	EF0576		Cationic ABC transporter, permease protein
ef0577	EF0577		Adhesion lipoprotein
ef0578	EF0578		Helix-turn-helix, iron-dependent repressor family
ef0579	EF0579	TAGACTCATCTAAA	Transcriptional regulator, putative
mntH2	EF1057	TAGGTGTACCTAAA	Mn2+/Fe2+ transporter, NRAMP family
ef1058	EF1058		Universal stress protein family
mntH1	EF1901	TAGGTGTGCCTAAA	Manganese transport protein MntH
efaC	EF2074	TAGGTGCGCCTAAA	ABC transporter, ATP-binding protein
efaB	EF2075		ABC transporter, permease protein
efaA	EF2076		Endocarditis-specific antigen

^aData are from reference 16.

^bGenes whose upstream (promoter) regions contain an EBM are in bold.

^cAccording to the National Center for Biotechnology Information (NCBIhttp://www.ncbi.nlm.nih.gov/).

Fig 2.

Effect of metal addition on E. faecalis. All strains carry plasmid pSAVE11, which contains a fusion of E. coli lacZ to the EBM-containing promoter of mntH2. (A) lacZ expression in wild-type VE14089 (pSAVE11) grown in GM17 with the indicated added concentrations of ZnSO₄, MnCl₂, CuSO₄, or FeCl₂. (B) lacZ expression in the wild type (pSAVE11) and the efaR mutant (pSAVE11) grown in GM17 with ZnSO₄ (0 mM or 4 mM), MnCl₂ (0 mM or 0.4 mM), or CuSO₄ (0 mM or 0.05 mM). *, P < 0.03.

EfaR regulates genes with an EBM.

Five of the genes/operons that contain an EBM in their promoter regions were previously shown to be upregulated in the presence of a high concentration of Zn²⁺ and downregulated with high concentrations of Mn²⁺ and Cu²⁺ ions (16). These genes and their operons, presented in Table 3, are mostly (predicted to be) involved in metal transport and regulation and include the efaCBA operon (ef2074 to ef2076) and the cluster ef0575 to ef0578, which is present in the pathogenicity island of E. faecalis V583. The genes ef0575 to ef0577 encode homologs of the EfaCBA proteins (identities of 52%, 60%, and 60%, respectively). The genes in Table 3 also include a putative TetR family regulator gene (ef0579) as well as mntH1 (ef1901) and mntH2 (ef1057), both annotated as being putatively involved in manganese transport, and a universal stress protein (USP) gene, ef1058, whose function and role in metal homeostasis, if any, have not been studied yet in the genus Enterococcus.

EfaR from JH2-2 strain binds to an EBM-containing promoter region, that of the efaCBA operon, in a manganese-dependent way (12). We wondered whether E. faecalis V583 EfaR, which is identical to EfaR of strain JH2-2, regulates the expression not only of efaCBA but also of the other genes identified here as containing an EBM site in their promoter regions. Thus, a single-crossover mutant of efaR was made in E. faecalis VE14089. DNA microarrays were performed to compare the expression of the genomes of VE14089 and its isogenic efaR mutant. Some genes with a copy of EBM in their promoter regions, namely, those listed in Table 3, were differentially expressed in the efaR mutant, although the P values are rather high (between 0.5 and 0.01) (see Table S2 in the supplemental material). Nevertheless, a hypergeometrical distribution test performed on these operons showed that they were highly significantly changed (28) (see Table S2). We chose one of the genes that was shown to be relevant after this test as a proof of principle that EfaR regulates genes with an EBM. Thus, to confirm that expression of mntH2 (ef1057) is dependent on EfaR, β-galactosidase activities driven by the P_mntH2-lacZ transcriptional fusion on plasmid pSAVE11 were compared in E. faecalis VE14089 and its efaR mutant (Fig. 2B). In the absence of added zinc, manganese, or copper ions, β-galactosidase activity was lower in the mutant than in the parent strain, suggesting that EfaR has a role in mntH2 transcription. When zinc, manganese, or copper ions were added to the efaR mutant carrying pSAVE11, no significant induction or repression of P_mntH2-lacZ was seen, as opposed to the response in the parent strain, VE14089(pSAVE11). These results suggest that the three metal ions function as effectors of EfaR regulation of mntH2 transcription and confirm the metal-dependent regulation of EBM-containing genes by EfaR.

To examine the effect of each metal ion individually, the same assays were performed in metal-depleted medium (Fig. 3). Under these conditions, a huge increase in P_mntH2-driven LacZ activity was observed in both the wild type and the efaR strain: lacZ expression was much lower under metal-replete conditions (Fig. 2A). Mn²⁺ was still able to repress P_mntH2-driven expression in the efaR mutant in the metal depleted environment (Fig. 3), suggesting that under such conditions yet another regulator is likely to be involved. Since one of the genes in Table 3, ef0579, encodes a putative regulator, we tested its role in metal sensing. To this end, the activities of the P_mntH2::lacZ and P_ef0575::lacZ fusions were measured in the absence of EF0579, using an insertional mutant of ef0579 (19). The results presented in Fig. 4A and B clearly show that EF0579 does not play a role in the metal-dependent regulation of ef0575 or mntH2, as the β-galactosidase activities in the wild-type strain and the ef0579 mutant were similar.

Fig 3.

Effect of metal addition on E. faecalis lacZ expression in the wild-type strain VE14089 and its isogenic efaR mutant, both carrying pSAVE11 with the P_mntH2::lacZ fusion, under metal-depleted conditions. Values are lacZ expression in the strains grown in chelGM17 with ZnSO₄ (0 mM or 4 mM), MnCl₂ (0 mM or 0.4 mM), or CuSO₄ (0 mM or 0.05 mM). *, P < 0.03.

Fig 4.

Effect of metal addition in wild-type E. faecalis VE14089 or its ef0579 mutant, each carrying pSAVE10, a plasmid with a Pef0575::lacZ fusion, or pSAVE11, a plasmid with a PmntH2::lacZ fusion. (A) lacZ expression in VE14089 (pSAVE10) and the ef0579 mutant (pSAVE10) grown in GM17 with ZnSO₄ (Zn; 0 and 4 mM), MnCl₂ (Mn; 0 and 0.4 mM), or CuSO₄ (Cu; 0 and 0.05 mM); (B) lacZ expression in VE14089 (pSAVE11) and the ef0579 mutant (pSAVE11) grown in GM17 with ZnSO₄ (Zn; 0 and 4 mM), MnCl₂ (Mn; 0 and 0.4 mM), or CuSO₄ (Cu; 0 and 0.05 mM). *, P < 0.03.

EfaR contributes to E. faecalis virulence.

As the results presented above show that EfaR is a relevant player in metal-dependent gene regulation in E. faecalis, we studied its further role(s) in the biology of this bacterium. Several EfaR homologues have been implicated in processes such as virulence, survival in macrophages, biofilm formation, and oxidative stress in Mycobacterium tuberculosis, Streptococcus gordonii, Streptococcus pneumoniae, and Streptococcus mutans (10, 29–31). The E. faecalis efaR mutant and several isogenic strains carrying mutations in genes regulated by EfaR were examined with respect to their capacity to survive inside macrophages, to form biofilms, and to stand oxidative stress.

As shown in Fig. 5, the E. faecalis efaR mutant was significantly impaired in its ability to survive inside macrophages compared to the wild-type strain VE14089 (P value of 4.24 × 10⁻⁵). The latter was even able to divide to some extent inside the macrophages for the first 2 h and only later succumbed to macrophage defenses. After 4 h postinfection, more than 90% of the efaR cells were eliminated by the macrophages, while during the same period of infection, wild-type E. faecalis VE14089 cells remained inside macrophages without dying. The huge impact of the efaR deletion on the ability of E. faecalis to survive inside macrophages suggests that it is involved in escaping host defenses, and thus, EfaR contributes to virulence of this species. In fact, the longer cells survive inside macrophages, the more likely they are to reach internal host sites and, consequently, to be successful in infection. The other mutants tested in this assay, namely, those affected in ef0577, ef0579, and efaA, behaved like the wild-type strain with respect to intramacrophage survival (data not shown). This result suggests that these genes play no direct role in escaping this type of host defense.

Fig 5.

Intramacrophage survival of E. faecalis. Percentage of survival of E. faecalis VE14089 (gray bars) and the efaR mutant (black bars) inside macrophages (J774 A1 cells) at the indicated times after addition of the bacteria to a confluent layer of macrophages. Bacterial survival was measured by plating on BHI agar plates. Results are the means from three independent experiments. *, P < 0.03, calculated in comparison with the value at time zero.

The ability to withstand oxidative stress is important for the success of infectious agents such as E. faecalis, as host antibacterial defenses often rely on the deleterious effects of reactive oxygen species, e.g., in macrophages. The tolerance of the various E. faecalis strains used in this study to oxidative stress was assessed by measuring survival of cells upon addition of a high concentration of hydrogen peroxide. This assay has been used to detect and characterize the hypR gene, which codes for a regulator of genes involved in the E. faecalis response to oxidative stress (25). Among all mutants tested, only the efaR mutant showed a decreased capacity to tolerate the oxidative stress imposed by hydrogen peroxide (Fig. 6). After 15 and 30 min of incubation, the efaR mutant was 31- and 38-fold, respectively, more sensitive to hydrogen peroxide than the wild type.

Fig 6.

E. faecalis tolerance to oxidative stress. Survival of E. faecalis VE14089 (diamonds) and its isogenic efaA, efaR, and ef0577 insertion mutants at 15, 30, and 60 min after challenge with 20 mM H₂O₂. CFU were determined by plate counting using BHI plates. A value of 100% corresponds to the number of CFU immediately prior to the additions of H₂O₂ (0 min). The values are the averages from three independent experiments. *, P < 0.03, calculated in comparison with the value at time zero.

Biofilms also contribute to the success of E. faecalis as an infectious agent, namely, during endocarditis, the type of infection that enabled the identification of EfaA as a virulence factor (13, 32). We thus assessed and compared the ability of VE14089 and several of its isogenic mutants to form biofilms by measuring biofilms formed overnight on the surfaces of polystyrene plates. EF0577 has 60% similarity to EfaA; both proteins have the same domain organization, i.e., they both putatively face the outside of the cell and are anchored to the cytoplasmic membrane through a transmembrane domain. Moreover, the genes of both proteins carry an EBM in their promoters, are regulated by EfaR in a manganese-dependent way, and are integrated into operons coding for manganese transporters. As neither of the mutants in the genes for these proteins was significantly impaired in the ability to form biofilms, both proteins seem not to be involved in biofilm formation in E. faecalis, although we cannot rule out the possibility that they may complement each other. Only the efaR mutant was significantly impaired in biofilm-forming ability (P values of 0.014) in comparison to its parent strain, VE14089 (Fig. 7). Indeed, the efaR mutant appeared to have completely lost the ability to form biofilms.

Fig 7.

E. faecalis biofilm formation. The biofilm-forming ability of wild-type E. faecalis VE14089 and its efaR, efaA, and ef0577 mutants was quantified using crystal violet staining (34). Values are averages of results obtained in three independent experiments, performed in six replicates. *, P < 0.03.

DISCUSSION

In this work we establish that EfaR is a major regulator in E. faecalis; it modulates the expression of several Mn²⁺-dependent systems using Mn²⁺ as a cofactor. We show that the operon ef0575-ef0578, of which the products of the first three genes are homologous to those of the efaCBA operon, is part of the EfaR regulon. Twenty-one of 33 E. faecalis genome sequences, including hospital and food strains, contain both efaCBA and ef0575-ef0578 (data not shown). Both operons respond to metals in a similar way. They were downregulated in the efaR mutant and both carry the EBM site in their promoter regions. Although the operons are annotated as manganese transport systems, their function as such remains to be demonstrated. Despite these similarities, there are relevant differences between the two, which suggests that they might play different roles in E. faecalis biology and virulence. For example, EF0578, for which no homologue is present in the efaCBA operon, is a putative small iron-dependent repressor with homology to part of EfaR and its DtxR-type homologs. Another difference between the efCBA and ef0575-ef0578 products concerns the EF0577 and EfaA proteins. Despite their amino acid sequence similarity, EF0577 and EfaA had different impacts on the ability of E. faecalis to form biofilms.

Previous DNA microarray results (16), the presence in their upstream sequences of the EBM site, and their annotation as Mn²⁺ transporter genes all suggest that the mntH genes are part of the EfaR regulon. Using the P_mntH2-lacZ fusion, we demonstrated that EfaR indeed has a regulatory role in mntH2 transcription, particularly under metal-replete conditions. It is also clear from our experiments that under metal-depleted conditions, transcription of mntH2 is much higher than in metal-replete medium. Most likely, the genes in Table 3 are expressed with the purpose of scavenging manganese ions in environments where this metal is extremely scarce. In Lactobacillus plantarum, the genes of the homologs of E. faecalis MntH1 and MntH2 were expressed upon Mn²⁺ limitation, supporting the role of both proteins in Mn²⁺ homeostasis (33). Although the role of these transporters in E. faecalis has not been studied, the fact that they are part of the EfaR regulon reinforces the theory that they are involved in manganese transport.

Iron was found to induce mntH2 transcription, suggesting that MntH2, which is annotated as a manganese/iron transporter, may also be involved in iron transport. Previous work by others (12) showed that the efaCBA promoter, containing an EBM, does not respond to iron, and that EfaR binding to the promoter region of efaCBA was not iron-dependent. The mntH2 promoter carries an EBM and also responds to iron. It is possible that it does so through another regulator, not via EfaR/EBM. The mntH2 gene is in an operon with ef1058, a gene annotated as encoding a universal stress protein (USP) and which behaved as mntH2 in our previous DNA microarray work (16). Several USPs of E. coli have been proven to be involved in iron scavenging, control of intracellular iron levels, and defense against reactive oxygen species (34). It is possible that the induction by iron of the mntH2-ef1058 operon serves to protect E. faecalis against the deleterious effects of excess iron.

In the presence of a low concentration of metals (in chelated medium), another regulator seems to be operative in E. faecalis, whose action was visible only in the absence of EfaR and in metal-depleted medium (Fig. 3). The genome of E. faecalis V583 specifies several predicted Mn²⁺ transport systems. It is likely that more than one regulator would affect the regulation of these different systems. Several genes with a putative EBM in their promoter region did not respond to the deletion of efaR in the microarray experiment under the conditions tested. This suggests either that these EBM sites are not genuine, that the genes react to different circumstances, or that they are regulated by a different regulator also capable of binding to (a variation of) the EBM. These possibilities remain to be explored. Although EfaR has been shown here to have a major metalloregulatory role, in environments as complex as the ones found in the host, manganese regulation in E. faecalis must be fine-tuned to be able to quickly adjust to changing conditions. Thus, it is likely that other players are involved, complementing the major function of EfaR in the maintenance of manganese homeostasis. The putative regulator EF0579 was ruled out for such a role, at least with respect to controlling ef0575 and mntH2 expression. Another candidate regulator would be EF0578, as it is also part of the EfaR regulon and shares 45% homology with EfaR. However, none of the residues that probably constitute the metal binding site in EfaR, by homology with the residues described for ScaR (35) discussed below, are present in EF0578, and thus, its function as a regulator has yet to be proven.

Structural information on EfaR is necessary to fully understand the interaction of metal ions with this member of the DtxR/MntR family of metalloregulators. A phylogenetic tree of these regulators indicates that ScaR from S. gordonii and EfaR from E. faecalis belong to the same group, sharing 41% identity (35). Based on structural studies of ScaR (35), it seems that the key residues involved in metal binding are also present in EfaR. ScaR responds selectively to manganese in vivo, despite being activated in vitro by other metal ions as well. Metal binding to metalloregulatory proteins is determined by affinity, allostery, and access (36). On the basis of the similarities in the (putative) metal binding sites in ScaR and EfaR, we presume that metal binding in both proteins occurs basically in the same way, with the manganese ion being the main coregulator. Zn²⁺ and Cu²⁺ do not seem to be true coregulators, as they do not affect expression from an EfaR-regulated promoter in metal-depleted medium (Fig. 3). Nevertheless, it has been shown that EfaR binds to the efaCBA promoter region in the presence of Mn²⁺, Zn²⁺, and Cu²⁺ (12), suggesting that Zn²⁺ and Cu²⁺ can compete with Mn²⁺ for EfaR binding, impairing manganese corepressor activity. Metal ions competing with each other for ligation to EfaR could strongly influence the maintenance of manganese homeostasis, which is essential for colonization and infection (see below).

Manganese has a major role in several cellular processes, among which is the response to oxidative stress. Destabilization of manganese homeostasis, in particular by the inability to scavenge this metal ion, could seriously affect a number of cellular processes and jeopardize the ability of a bacterium to survive in and colonize its host. As a key player in manganese-dependent gene expression in E. faecalis, EfaR could be involved in processes relevant for host-pathogen interactions. In fact, it has been observed in other pathogens that mutations in EfaR homologs affect biofilm formation (SloR) (29), intramacrophage survival (IdeR) (30) and oxidative stress tolerance (ScaR) (31). Indeed, a mutant of EfaR was greatly impaired in the ability to survive inside macrophages and to tolerate oxidative stress. Macrophages use reactive oxygen species to kill invading pathogens, and it is highly likely that the decreased tolerance to oxidative stress of the efaR mutant accounts for its inability to survive inside macrophages. Although the immediate and exact effects of hydrogen peroxide on gene expression and protein damage in E. faecalis have not been thoroughly studied, several enzymes and regulators are known to be involved in resistance to this oxidant (37). In E. coli, it is known that the pentose-phosphate pathway is negatively affected if enzymes involved in protection against hydrogen peroxide are absent (38). The latter study also revealed that hydrogen peroxide induces a shift from using manganese instead of iron as a cofactor in certain proteins, with increased manganese import and iron sequestration. It is possible that the decreased survival of the E. faecalis efaR mutant upon hydrogen peroxide treatment and inside macrophages is due to the inability of this strain to make such a shift from iron to manganese usage. EfaR was here identified as a major player in manganese homeostasis by controlling several putative manganese and iron transporters; its mutation led to repression of the sodA gene (change, 6.1-fold; P = 2 × 10⁻⁵) and efaCBA and ef0575-ef0578 transporter genes (see Table S2 in the supplemental material). The S. aureus EfaR homologue MntR regulates expression of manganese uptake systems and was shown to modulate PerR-regulated genes through control of manganese uptake (39). It is possible that EfaR plays the same role in E. faecalis with regard to HypR-regulated genes, as we found that sodA (this work) and ahpC (16) were differentially expressed in the efaR mutant and in the presence of excess manganese, respectively. Future studies on E. faecalis response to hydrogen peroxide should establish the network of events set in motion by this oxidant, but it appears that EfaR, through manganese homeostasis control, is a key player. Other genes whose expression directly or indirectly depends on the presence of a functional EfaR also contribute to the very short life of the efaR mutant inside macrophages. Although we do not know which E. faecalis genes are induced upon entry into macrophages, it is known that the SOS response is induced in some bacteria upon macrophage entry (40). In particular, the single-strand-binding protein Ssb1, which plays a role in DNA replication, repair, and recombination, is part of the SOS response (41). The repression of ssb1 (change, 7.8-fold; P = 2.87 × 10⁻⁶) in the efaR mutant may deprive this strain of one mechanism to properly repair DNA damage imposed by macrophage defenses and suggests that a link exists between EfaR and the E. faecalis SOS response.

EfaR impairment led to a lower ability of E. faecalis to form biofilms. Although alterations in biofilm architecture have been described for an S. mutans sloR mutant (29), this is the first description of a substantial decrease in biofilm formation by a DtxR/MntR family regulator mutant. The effect of the efaR mutation surpassed that of the genes under its regulation, namely, efaA and ef0577. These findings suggest that more EfaR-regulated genes are, together, responsible for the loss of biofilm-forming ability in the regulator mutant. One candidate is glmU (repressed in the efaR mutant with a change of 3.2-fold; P = 6.03 × 10⁻⁵). GlmU is involved in cell wall synthesis, and inhibitors of GlmU were found to nearly abolish biofilm formation by S. epidermidis (42).

Overall, this work highlights the importance of metal homeostasis in the regulation of processes important for E. faecalis virulence and the role of EfaR therein. In particular, manganese appears to be a key player in the regulation of EfaR activity. However, other metal ions most likely also contribute to E. faecalis virulence. In the case of zinc, this is probably through competition with manganese for EfaR binding. We show that EfaR is an important regulator of biofilm formation, intramacrophage survival, and oxidative stress tolerance. This is the first description of a DtxR/MntR regulator whose absence has such significant implications in different cellular processes crucial for survival and host colonization. The ability of EfaR to sense and control intracellular metal ion concentrations could also indirectly affect the activity, not the expression, of other metal-responsive regulators that are involved in induction and repression of oxidative stress response genes, metabolic enzyme genes, and DNA repair enzyme genes. From a broader perspective, this work identifies EfaR as a major regulator in E. faecalis and, thus, a target to be considered for therapeutic intervention against enterococcal infections.