Involvement of Peptidylprolyl cis/trans Isomerases in Enterococcus faecalis Virulence

Abstract

Peptidylprolyl cis/trans isomerases (PPIases) are enzymes involved in protein folding. Analysis of the genome sequence of Enterococcus faecalis V583 allowed for identification of 3 PPIases carrying genes. ef2898 encodes an intracellular PPIase which was not shown to be important for the E. faecalis stress response or virulence. The other two PPIases, the parvulin family rotamase EF0685 and the cyclophilin family member EF1534, are expected to be surface-exposed proteins. They were shown to be important for virulence and resistance to NaCl. A Δef0685 Δef1534 mutant was also more resistant to oxidative stress, was able to grow under a high manganese concentration, and showed altered resistance to ampicillin and quinolone antibiotics.

INTRODUCTION

Peptidylprolyl cis/trans isomerases (PPIases; also named foldases or maturases) are ubiquitous folding proteins (present in eukaryote or prokaryote cells) catalyzing the cis/trans isomerization of peptidylprolyl. They can be intra- or extracellular and are divided into three typical foldase families, including the cyclophilins, which are able to bind the immunosuppressant cyclosporine, the FK506 drug-binding proteins (FKBP), and the parvulin family, which refers to the described Escherichia coli parvulin (13, 29). These foldases ensure the conformation of proteins involved in many vital functions but also in virulence. In Gram-negative bacteria, SurA is a parvulin periplasmic folding factor involved in maturation of outer membrane proteins. SurA is also important for adhesion and invasion of E. coli, Shigella flexneri, and Salmonella enterica serovar Typhimurium (4). This protein is also involved in the oxidative stress resistance and virulence of Yersinia pseudotuberculosis (27). The macrophage infectivity potentiator (MIP) is a FKBP foldase necessary for the adhesion to collagen, and its PPIase activity renders the extracellular matrix sensitive to serine protease, which allows the invasion of Legionella pneumophila through the epithelial barrier (15). In Gram-positive bacteria, PPIases were also involved in bacterial virulence. For example, the maturase lipoprotein of the parvulin family PrtM was shown to be important for Streptococcus equi colonization in mouse and pony models (12). The parvulin foldase PrsA is essential for Bacillus subtilis viability (16) and required for protein secretion and production of fully mature streptococcal pyogenic exotoxin B (SpeB) in group A Streptococcus (22). In Listeria monocytogenes, the lipoprotein chaperone PrsA2 is involved in virulence factor secretion and viability within host cell cytosol and is regulated by the regulator of virulence factors PrfA (1). Among the four PPIases identified in Streptococcus pneumoniae, two are lipoproteins and, consequently, surface located. One of them, the rotamase lipoprotein SlrA (streptococcal rotamase A), belongs to the cyclophilin family, while the second one, PpmA (proteinase maturation protein A), shares homologies to PPIases of the parvulin family. Both are involved in the colonization phase of infection and in the evasion of phagocytes. The construction of an slrA ppmA double mutant showed a phenotype of reduced virulence slightly more important than those of the corresponding single mutants (7, 13, 28). These examples highlight the importance of PPIases from Gram-positive bacteria in virulence.

Enterococcus faecalis is an opportunistic bacterium normally present in the human gastrointestinal tract which became an emergent pathogen responsible for nosocomial infections in many hospitals in the United States and Europe (26). Involvement of 3 E. faecalis PPIases in stress response or in virulence was investigated here. Among them, 2 of these (EF0685 and EF1534) are lipoproteins anchored to the membrane and are predisposed to be involved in the folding of secreted proteins.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

The E. faecalis V19 parental strain (Table 1) used in this study is a plasmid-cured derivative of the vancomycin-resistant clinical isolate V583 (38). E. faecalis V19 and its derivates were cultivated at 37°C in M17 medium (36) supplemented with 0.5% glucose (GM17) or in M17 MOPS (morpholinepropanesulfonic acid) medium supplemented with glycerol as the sole carbohydrate (17). For growth under a high concentration of manganese, a chemically defined medium, MCDE Caa Trp medium (32), was used. E. coli XL1-Blue (Table 1) was used as the recipient for cloning. E. coli strains were cultivated under conditions of vigorous agitation at 37°C in Luria Bertani medium (33). When required, ampicillin (100 μg/ml) or erythromycin (150 μg/ml) was added to bacterial cultures.

Table 1.

Bacterial strains and plasmids

Strain or plasmid	Relevant property(ies)	Reference or source
E. faecalis strains
V19	Plasmid-cured strain obtained from E. faecalis V583	38
Δef0685	ef0685 deletion mutant obtained from V19	This study
Δef0685C	Δef0685 mutant derivative complemented for ef0685	This study
Δef1534	ef1534 deletion mutant obtained from V19	This study
Δef1534C	Δef1534 mutant derivative complemented for ef1534	This study
Δef0685 Δef1534	Mutant with deletions in ef0685 and ef1534 obtained from V19	This study
Δef1534 Δef2898	Mutant with deletions in ef1534 and ef2898 obtained from V19	This study
Δef1534C Δef2898	Δef1534 Δef2898 derivative complemented for ef1534	This study
Δef1534 Δef2898C	Δef1534 Δef2898 derivative complemented for ef2898	This study
E. coli strain XL1-Blue	recA1 endA1 gyrA96 thi-1 hsdR17 supE44 relA1 lac[F′ proAB lacIqZΔM15] Tn10 (Tetr)	Stratagene
Plasmids
pGEMt	System for cloning PCR products	Promega
pGEMtwtef0685	pGEMt derivative carrying a 4.0-kb DNA fragment from ef0685	This study
pGEMtwtef1534	pGEMt derivative carrying a 2.9-kb DNA fragment from ef1534	This study
pGEMtwtef2898	pGEMt derivative carrying a 3.3-kb DNA fragment from ef2898	This study
pGHost9	Plasmid with thermosensitive replication	24
pGHost9wtef0685	pGHost9 derivative carrying a 4.0-kb DNA fragment from ef0685	This study
pGHost9wtef1534	pGHost9 derivative carrying a 2.9-kb DNA fragment from ef1534	This study
pGHost9wtef2898	pGHost9 derivative carrying a 3.3-kb DNA fragment from ef2898	This study
pGHost9Δef0685	pGHost9 derivative carrying a 3.1-kb DNA fragment with a deleted ef0685 gene	This study
pGHost9Δef1534	pGHost9 derivative carrying a 2.3-kb DNA fragment with a deleted ef1534 gene	This study
pGHost9Δef2898	pGHost9 derivative carrying a 2.8-kb DNA fragment with a deleted ef2898 gene	This study

General molecular methods.

Genomic DNA extraction was performed as described by Sambrook et al. (33). GoTaq Flexi DNA polymerase (Promega, Madison, WI) was used for PCRs with the primers listed in Table 2. PCR products were purified using the NucleoSpin extract II kit (Macherey-Nagel, Dünen, Germany). Plasmids used in this study are listed in Table 1. They were extracted with the NucleoSpin plasmid kit (Macherey-Nagel). Restriction endonucleases and T4 DNA ligase were purchased from Promega and used according to the manufacturer's instructions.

Table 2.

Primers used in this study

Primer name	Sequence (5′–3′)a
Ef0685m0	GCAGGAAACCATTCCAATTT
Ef0685m1	TACGAGCTCGTTTGAACGCTTCTT (SacI)
Ef0685m2	ACTGGATCCGCTGCAGCTAAGATT (BamHI)
Ef0685m3	TTCGGATCCCAAAACAAGCGACTC (BamHI)
Ef0685m4	TCTGCATGCACTAGCTCAACGTTGGTCT (SphI)
Ef0685m5	ACCTTAGATGAGTTGTATCAA
Ef1534m0	TGAGACAATTTTGTTGCGAC
Ef1534m1	CACGAATTCTACCTCTTGCGCTATGCTCC (EcoRI)
Ef1534m2	TCAGGATCCGTTCCCCTCGTGCCTTTA (BamHI)
Ef1534m3	CAAGGATCCAAAGTCGAAGTTGGTGCC (BamHI)
Ef1534m4	AATACTGCAGATGTCGGTGCCTCAG (PstI)
Ef1534m5	AAATCGTTATTTCATCGTATGC
Ef2898m0	GATGATAAATGATCGGTAGATG
Ef2898m1	TTCTGTCGACCCACAGCTTTCACTTGTTCC (SalI)
Ef2898m2	TCAGGATCCGCTTTTGGGCCTTTAGC (BamHI)
Ef2898m3	CAAGGATCCGTCCAACATTTGACGTCG (BamHI)
Ef2898m4	AATTCCATGGCGTATACAGCACCTGCTG (NcoI)
Ef2898m5	GAAATTTTGTTTAATAGTGGCG
PU	TGTAAAACGACGGCCAGT
PR	CAGGAAACAGCTATGACC

^aThe underlined sequences correspond to the recognition sites of the restriction endonucleases shown in parentheses.

Construction of mutants affected in ef0685, ef1534, and ef2898 and complementation.

For the construction of an ef0685 deletion mutant (Δef0685), allelic replacement was used as described below. Ninety percent of the ef0685 gene was deleted in the mutant strain, starting 27 bp after the ATG start codon and stopping 65 bp before the TAA stop codon. Briefly, a 4.0-kb DNA fragment (obtained using chromosomal DNA of E. faecalis V19 as the template) containing the ef0685 gene flanked by 1,486 bp upstream and 1,456 bp downstream was amplified using primers Ef0685m1 and Ef0685m4 (Table 2) and cloned into the pGEMt plasmid (Table 1) to obtain pGEMtwtef0685. Then, an inverse PCR was done with primers Ef0685m2 and Ef0685m3 (Table 2) to allow a 930-bp deletion. After digestion with BamHI, the fragment was recircularized with T₄ DNA ligase. The common SacII restriction site between the pGEMt and pGHost9 plasmids (Table 1) was used to add pGHost9 to the previous construction. The next step was a digestion with the restriction endonuclease NotI to remove the pGEMt vector and to obtain a plasmid composed of pGHost9 with the deleted fragment of ef0685 (pGHost9Δef0685). The corresponding recombinant plasmid was then introduced in E. faecalis V19. Transformants obtained at 30°C (erythromycin-resistant colonies) were used for temperature shifts in order to select clones with the pGHost9Δef0685 plasmid integrated within the chromosome (first crossing-over) and then clones resulting from a double crossing-over. These were analyzed for the presence of a deleted ef0685 gene, and the corresponding region was verified by sequencing.

For the construction of the ef1534 and ef2898 deletion mutants (Δef1534 and Δef2898, respectively), the same strategy was used. Primers Ef1534m1 and Ef1534m4 for ef1534 and Ef2898m1 and Ef2898m4 for ef2898 (Table 2) allowed for amplification of 2.9-kb and 3.3-kb DNA fragments, respectively, which were cloned in pGEMt to produce pGEMtwtef1534 and pGEMtwtef2898 (Table 1). Inverse PCRs were performed with oligonucleotides Ef1534m2/Ef1534m3 and Ef2898m2/Ef2898m3 to generate 565-bp and 467-bp deletions for ef1534 and ef2898, respectively. With the method described above, pGHost9 was added to these constructions, and then pGEMt was removed to obtain pGHost9Δef1534 and pGHost9Δef2898. The corresponding recombinant plasmids were introduced in E. faecalis V19, and temperature shifts allowed mutants to be obtained. These were analyzed for the presence of a deleted gene, and the corresponding region was verified by sequencing. We constructed the Δef0685 Δef1534 and Δef1534 Δef2898 double mutants using the Δef0685 strain and pGHost9Δef1534 and Δef1534 and pGHost9Δef2898, respectively.

To complement the ef0685 mutation, the 4.0-kb PCR fragment containing the wild-type (wt) ef0685 gene (Ef0685m1-Ef0685m4) was inserted into pGHost9 to obtain pGHost9wtef0685. This plasmid was then introduced into the Δef0685 mutant, and double crossover events allowed the construction of the complemented Δef0685C strain. The same strategy was used to complement the Δef1534 and Δef2898 mutants with plasmids pGHost9wtef1534 and pGHost9wtef2898, respectively. All complemented strains were analyzed by PCR, and the corresponding region was verified by sequencing.

Survival experiments.

E. faecalis survival to sodium chloride (28.5%), bile salts (0.3%), ethanol (22%), heat (62°C), and incubation in urine from humans (healthy volunteers) was tested as described by Reffuveille et al. (32). Susceptibility of E. faecalis strains to antibiotics was estimated using Mueller-Hinton medium plates with discs containing ampicillin (10 μg), penicillin (6 μg), gentamicin (500 μg), kanamycin (1 mg), netilmicin (30 μg), streptomycin (500 μg), tobramycin (10 μg), doxycycline (30 μg), minocycline (30 IU), tetracycline (30 μg), ciprofloxacin (5 μg), levofloxacin (5 μg), nalidixic acid (30 μg), norfloxacin (5 μg), ofloxacin (5 μg), teicoplanin (30 μg), or vancomycin (30 μg).

Detection of SOD activity.

Bacteria from exponential growth phase in GM17 medium (optical density at 600 nm [OD₆₀₀] of 0.5) were exposed to 1.5 mM H₂O₂ for 30 min. Total proteins were then extracted, and 15 μg was separated by nondenaturing PAGE (carried out according to the protocol of Laemmli [18], except that SDS and β-mercaptoethanol were omitted). The superoxide dismutase (SOD) activity was detected using the protocol of Beauchamp and Fridovich (3).

Infection and survival experiments of Galleria mellonella.

Infection of G. mellonella larvae with E. faecalis was accomplished as previously described by Lebreton et al. (19). Briefly, using a syringe pump (KD Scientific, Holliston, MA), larvae (about 0.3 g and 3 cm in length) were infected subcutaneously with washed E. faecalis cells from an overnight culture in GM17, with 1.5 × 10⁶ ± 0.15 × 10⁶ or 3 × 10⁶ ± 0.3 × 10⁶ CFU per larva administered in 10 μl of sterile saline buffer. In each test, 15 insects were infected, and the experiments were repeated at least three times. Larval killing was then monitored each hour from 18 h to 24 h postinfection.

Caco-2/TC7 culture and infection.

Caco-2/TC7 cells were routinely grown in Dulbecco's modified Eagle's medium (DMEM; Invitrogen) supplemented with 15% heat-inactivated fetal calf serum, 2 mM l-glutamine, 100 U/ml each of penicillin and streptomycin, and 1% nonessential amino acids. For experimental assays, the cells were seeded at a density of approximately 1 × 10⁵ cells/cm² in 24-wells tissue culture plates cultured at 37°C in 5% CO₂-95% air atmosphere, and the medium was changed daily. For infection, the E. faecalis strains in early stationary phase (grown overnight) were harvested by centrifugation (5,000 × g, 5 min, 20°C) and resuspended at a density of 10⁸ bacteria/ml in culture medium without serum and antibiotics. Confluent Caco-2/TC7 cells were washed twice with fresh medium, and the bacterial suspensions were then applied at the surface of the cell monolayers.

Adhesion and cytotoxicity assays on Caco-2/TC7 cells.

For adhesion, Caco-2/TC7 cells were infected with the E. faecalis strains for 4 h. At the end of incubation, nonadherent bacteria were removed by PBS (phosphate-buffered saline), and cells were lysed for 15 min with 0.1% Triton X-100. The lysates were diluted and plated on agar plates to determine the number of adherent/invasive bacteria.

For the cytotoxicity assay, Caco-2/TC7 cells were infected with the E. faecalis strains for 24 h. At the end of incubation, the cytotoxicity of the bacterial strains was estimated using two complementary methods: the lysed cells were quantified by measuring the lactate dehydrogenase (LDH) released in the cell supernatants as described by Madi et al. (23), and the viability of the cells was performed using the neutral red (NR) assay according to the procedure of Fautz et al. (10). Briefly, the cells were stained with 0.2 ml of fresh medium containing 50 μg/ml of NR for 1 h (37°C, 5% CO₂, 95% relative humidity) and were rinsed once with PBS. The NR was extracted with 1% (vol/vol) acetic acid in a 1:1 water-ethanol mixture. The absorbance of the extracted dye was spectrophotometrically measured at 540 nm.

Systemic murine infection experiments.

Mouse infections were done as described by Michaux et al. (25). Briefly, bacteria were cultivated in brain heart infusion (BHI) medium with 40% heat-inactivated horse serum. Ten female BALB/c mice were infected through the tail vein with 100 μl of a PBS solution containing 10⁹ bacteria/ml. After 7 days of infection, mice were euthanized by CO₂ asphyxiation. Their kidneys and their liver were removed, weighed, and homogenized in 5 ml of PBS buffer. The quantity of bacteria was determined by serial dilutions on an Enterococcus agar plate.

RESULTS AND DISCUSSION

Identification and mutagenesis of genes encoding PPIases in E. faecalis.

To identify the genes encoding PPIases in E. faecalis V583, we searched for homologs to the characterized S. pneumoniae PpmA and SlrA PPIases (7, 13) using BLASTp software (http://blast.ncbi.nlm.nih.gov/Blast.cgi) (2). One E. faecalis V583 open reading frame (ORF) encodes a protein (EF0685) which exhibits 119/302 (39%) amino acid sequence identity and 178/302 (58%) similarity to the S. pneumoniae PpmA PPIase (Fig. 1A), while two genes encode SlrA homologs. Indeed, ef1534 encodes a protein which exhibits 116/238 (48%) identity and 150/238 (63%) similarity to SlrA of S. pneumoniae, and a protein deduced from ef2898 shows 94/204 (46%) identity and 131/204 (64%) similarity with this PPIase (Fig. 1B). The EF0685 and EF1534 proteins were previously predicted to be lipoproteins (31) and consequently are expected to act on folding and trafficking of extracellular proteins. EF1534 and EF2898 share a high level of homology (52% identity and 68% similarity [Fig. 1B]); however, no predicted peptide signal or transmembrane domain was found in EF2898, suggesting it could be an intracellular folding protein. The ef0685 ORF has been annotated as a rotamase of the parvulin family in the genome of the E. faecalis V583 strain. It encodes a protein of 342 amino acids with a calculated molecular mass of 37.4 kDa. Both ORFs ef1534 and ef2898 were annotated as PPIases, encoding a 27.4-kDa protein consisting of 249 amino acids and a 196-amino-acid-long protein of 21.5 kDa, respectively.

Fig 1.

Alignments of EF0685, EF1534, and EF2898 with PPIases from S. pneumoniae. (A) E. faecalis EF0685 was aligned with S. pneumoniae PpmA, which belongs to the parvulin family. (B) E. faecalis EF1534 and EF2898 were aligned with S. pneumoniae SlrA, which belongs to the cyclophilin A family. Amino acids constituting putative peptide signals were overlined or underlined, and cysteines expected to be located in position +1 of the mature lipoproteins are indicated by asterisks. Residues identical in 2 proteins are presented on a black background, and functionally related amino acids (H, K, and R; F, Y, and W; L, I, M, and V; G and A; S and T; D and E; N and Q; and C and K) are on a gray background.

Search for conserved domains using Pfam (http://pfam.sanger.ac.uk/search [11]) within the amino acid sequences of the 13 E. faecalis V583 ORFs assigned to the protein folding and stabilization role category (30) revealed the existence of a fourth PPIase. It (EF0715) contains the Pfam05697, Pfam00254, and Pfam05698 domains, which are characteristic of ribosome-associated trigger factors and consequently would be involved in the folding of newly synthesized proteins (8).

As several PPIases from Gram-positive bacteria were shown to be involved in virulence and particularly in host colonization, we investigated the role of the E. faecalis EF0685, EF1534, and EF2898 PPIases in stress resistance and virulence by construction of single Δef0685, Δef1534, and Δef2898 mutants as described in Materials and Methods. As EF1534 and EF2898 share high similarities, and because EF1534 and EF0685 were predicted to be lipoproteins and consequently are both expected to be involved in the folding of extracellular proteins, we also constructed the Δef1534 Δef2898 and Δef0685 Δef1534 double mutants (see Materials and Methods). Since the construction of mutants with deletions of 90%, 78%, and 80% of the ORFs ef0685, ef1534, and ef2898, respectively, was successful, none of these genes appeared essential for E. faecalis growth.

Phenotypic characterization of mutants.

When growth of the Δef0685, Δef1534, Δef2898, Δef1534 Δef2898, and Δef0685 Δef1534 mutants was compared to that of the wild-type strain using GM17 medium with or without shaking during incubation, no significant difference was observed (data not shown).

The behavior of the wild-type and mutant strains to sodium chloride (28.5%), bile salts (0.3%), urine, ethanol (22%), and antibiotics exposure at different pH (4.0 to 9.0) or at 62°C was assayed, as indicated in Materials and Methods. The viability of Δef0685 and Δef1534 was affected by high sodium chloride concentration, while the complemented strains restored the wild-type phenotype (Fig. 2). Under this condition, the viability of Δef0685 Δef1534 was even more impaired than that of the single mutants (Fig. 2). PPIases involvement in NaCl stress was already observed in Shewanella sp. strain WP3 through overexpression of a gene encoding a PPIase during growth in the presence of NaCl (20). NaCl is known to generate important damages to the folding of cell surface proteins (14, 37). Consequently, EF0685 and EF1534 may be involved in the repair of damages caused by a high NaCl concentration on the surface of bacterial cells. These extracellular foldases may also play a role in the folding of ABC transporter components like the lipoprotein encoded by ef0863, which corresponds to the substrate binding subunit of an osmolyte transporter (OpuAC). So the NaCl sensitivity of Δef0685 and Δef1534 could also be explained by a reduction of osmolyte intake due to misfolding of ABC transporter components in the absence of EF0685 and EF1534.

Fig 2.

Effect of a high NaCl concentration on survival of wild-type and mutant strains. Bacteria grown in GM17 medium to an OD₆₀₀ of 0.5 were harvested and resuspended in saline solution containing 28.5% of NaCl. The percent viability represents the ratio between the number of cells surviving after 24 h or 48 h and the number of cells prior to challenge. Results are those of the E. faecalis wild-type strain (black bars), Δef0685 and Δef1534 single mutants, Δef0685 Δef1534 double mutant (white bars), and Δef0685C and Δef1534C complemented strains (gray bars). Student's t test was used to determine whether results for the wild type and other strains were significantly different. *, P < 0.05.

Since E. faecalis lipoproteins were shown to be involved in oxidative stress response (32), we analyzed growth of mutants under oxidative stress conditions. No difference with the three single mutants or with the Δef1534 Δef2898 double mutant was observed when they were incubated in GM17 medium supplemented with 1.5 mM H₂O₂ (data not shown). Alternatively, the Δef0685 Δef1534 strain appeared to have an advantage under this condition (Fig. 3A). Growth of the mutants was also evaluated in CCM17 MOPS medium supplemented with 0.5% glycerol (a condition which allows H₂O₂ production within the cytoplasm [5]) (Fig. 3B). Under this condition again, growth of the Δef0685 Δef1534 double mutant was less affected than that of the wild-type strain or single mutants (OD₆₀₀ of 0.85 at 8 h for Δef0685 Δef1534 versus OD₆₀₀ of 0.23 for the wild-type strain). As lipoprotein PPIases act on the conformation of extracellular proteins, we can suggest that the wild-type strain or single mutants could be targets of oxidation or could enhance the effect of oxidative stress and that their misconformation contributes to the phenotype observed here in protecting cells of oxidation.

Fig 3.

Effect of ef0685 and ef1534 inactivation on growth. Growth of the E. faecalis wild type (filled diamonds) and Δef0685 Δef1534 mutant (open squares) was determined by optical density at 600 nm (OD₆₀₀). (A) Growth in GM17 medium supplemented with 1.5 mM H₂O₂ under shaking; (B) growth in CCM17 MOPS medium supplemented with 0.5% glycerol under shaking; (C) growth in MCDE Caa Trp medium with 1.76 mM manganese at 37°C with shaking.

Activity of the superoxide dismutase (SOD) was evaluated using zymographic analysis (Fig. 4) and revealed that the growth advantage for the Δef0685 Δef1534 double mutant under oxidative stress conditions is not due to a change in the level of the SOD activity.

Fig 4.

Activity of the superoxide dismutase enzyme detected by zymographic analysis. Proteins were extracted from the E. faecalis wild type, the Δef0685, Δef1534, and Δef0685 Δef1534 mutants, and the Δef0685C and Δef1534C complemented strains.

Antibiograms were performed to compare antibiotic resistance levels of the mutants and the wild-type strain. Areas of growth inhibition for the single Δef0685, Δef1534, and Δef2898 mutants and the Δef1534 Δef2898 mutant were similar to that of the wild-type strain (data not shown). However, important differences were observed when the Δef0685 Δef1534 double mutant was exposed to ampicillin, ofloxacin, norfloxacin, and ciprofloxacin (Table 3). Indeed, the Δef0685 Δef1534 mutant presented sensitivity toward ampicillin, while smaller areas of inhibition for ofloxacin, norfloxacin, and ciprofloxacin revealed resistances of the double mutant to these antibiotics from the quinolone family. Ampicillin acts by inhibiting cell wall synthesis. The inactivation of both extracellular chaperones EF0685 and EF1534 is expected to have an effect on conformation of secreted proteins, which could explain the phenotype of the Δef0685 Δef1534 double mutant to antibiotics interfering with the synthesis of the cell wall. The antibacterial action of quinolones is directed against the DNA replication with a second oxidative effect (9). The phenotypes of the Δef0685 Δef1534 double mutant in regard to ofloxacin, norfloxacin, and ciprofloxacin are probably the consequence of the difference in intake of these antibiotics due to the misconformation of extracellular proteins or could be linked to the increased ability of the corresponding mutant to grow in the presence of the oxidant.

Table 3.

Diameters of the inhibition areas (mm) observed with antibiogram analysis

Antibiotic	Diameter (mm) of inhibition area
	Wild type	Δef685 Δef2898 mutant
Ampicillin	14.5 (±1.8)	20.5 (±0.3)
Ofloxacin	22.5 (±0.3)	13.0 (±0.1)
Norfloxacin	23.0 (±0.1)	12.5 (±0.3)
Ciprofloxacin	29.0 (±0.1)	18.0 (±0.3)

As several membrane proteins are expected to be involved in the transport of components such as divalent transition metal ions, which are essential or in contrast could be toxic at high concentrations (21), we tested in a chemical defined medium the effect of a high manganese concentration on the growth of the wild-type strain and the mutants. No difference was observed for the Δef0685, Δef1534, and Δef2898 single mutants and the Δef1534 Δef2898 double mutant (data not shown). However, a clear advantage was seen for the Δef0685 Δef1534 double mutant for growth in the presence of manganese compared to the wild-type strain (final OD₆₀₀ of 2.4 for the mutant versus 0.7 for the wild-type strain) (Fig. 3C). This phenotype for Δef0685 Δef1534 is similar to that previously observed with an lgt mutant (32) and probably results in both cases from a decrease in Mn²⁺ internalization. This suggests that the absence of lipidated lipoproteins in the lgt mutant or the absence of the EF0685 rotamase and the EF1534 PPIase is probably responsible of the misconformation of proteins involved in the transport of manganese. The better growth of the Δef0685 Δef1534 double mutant under oxidative stress and its capacity to growth under a high manganese concentration are phenotypes similar to those previously observed for an lgt mutant (32). In this lgt mutant, the observed corresponding phenotypes could be due to nonlipidated EF0685 and EF1534, which probably became partially inefficient and led to misconformation of secreted or surface-exposed proteins.

In the presence of H₂O₂ or with a high Mn²⁺ concentration, the Δef0685 Δef1534 double mutant presented a clear advantage for growth compared to the wild-type strain, while the Δef0685 and Δef1534 single mutants were as sensitive as the wild-type strain. Also, in the presence of a high sodium chloride concentration, the survival of the double mutant was more affected than that of single mutants, and phenotypes faced to antibiotics exposure were visible only with Δef0685 Δef1534. Consequently, we can suggest that one PPIase might replace the absence of the other or if the two PPIases do not possess the same protein targets, targets of one PPIase might replace targets of the other.

Involvement of PPIase-encoding genes in E. faecalis fitness.

As described in the introduction, PPIases were shown to be involved in the fitness and, consequently, in the virulence of Gram-positive bacteria. In order to determine the involvement of the EF0685, EF1534, and EF2898 PPIases in virulence, Galleria mellonella caterpillars were infected with the wild-type strain or mutants (Δef0685, Δef1534, Δef2898, Δef1534 Δef2898, and Δef0685 Δef1534), as described in Materials and Methods. Larvae infected with Δef0685 at 3 × 10⁶ cells/larva survived bacterial infection better, as demonstrated by the survival rate monitored at 18 h and 20 h postinfection (Fig. 5A). On the other hand, the Δef1534 mutant showed a small attenuation of virulence only visible with 1.5 × 10⁶ cells/larva (Fig. 5B), while caterpillar survival after infection with Δef2898 was not significantly different from that observed with the wild-type strain (data not shown). The Δef0685 Δef1534 double mutant had a similar loss of virulence to Δef0685 (Fig. 5A). G. mellonella larvae were also infected with the Δef1534 Δef2898 double mutant and the Δef1534C Δef2898 (Fig. 5B) and Δef1534 Δef2898C (data not shown) complemented strains. The Δef1534 Δef2898 mutant presented attenuated virulence similar to that of the Δef1534 mutant, and when it was complemented with ef1534, bacteria recovered the wild-type phenotype (Fig. 5B, Δef1534C Δef2898). However, when the Δef1534 Δef2898 mutant was complemented with ef2898, no reduction of larval survival was observed compared to the double mutant (data not shown). Taken together, these data suggest that the extracytoplasmic foldases EF0685 and EF1534 are important for E. faecalis virulence but not the intracytoplasmic chaperone EF2898, and the involvement of the EF1534 PPIase was less important than that of the EF0685 rotamase. The involvement of extracellular PPIase in virulence is in accordance with results observed with the Gram-positive bacteria group A Streptococcus (22), S. equi (12), L. monocytogenes (1), and S. pneumoniae (7, 13, 28).

Fig 5.

Effect of PPIase inactivation on G. mellonella survival. Percent survival of G. mellonella from 18 h to 20 h postinfection with the E. faecalis wild type, mutants, and complemented strains. (A) A total of 3 × 10⁶ CFU counted on an agar plate of the E. faecalis wild type (black bars), Δef0685 and Δef0685 Δef1534 mutant strains (white bars), and Δef0685C complemented strain (gray bar) were injected per larva. (B) A total of 1.5 × 10⁶ CFU counted on an agar plate of the E. faecalis wild type (black bars), Δef1534 and Δef1534 Δef2898 mutant strains (white bars), and Δef1534C and Δef1534C Δef2898 complemented strains (gray bars) were injected per larva. Student's t test was used to determine whether results for the wild type and other strains were significantly different. *, P < 0.05.

Further experiments were performed to determine the effect of ef0685 and ef1534 inactivation on adhesion to eukaryotic cells, colonization, and cytotoxicity.

To test the capacity of bacteria to adhere to eukaryotic cells, we infected Caco-2/TC7 cells with the Δef0685, Δef1534, and wild-type strains. After 4 h of incubation, the number of CFU on agar plates were counted. No significant difference was observed in the adhesion capacities of the Δef0685 and Δef1534 mutants, which succeeded in eukaryotic cells adhesion as well as the wild-type strain (data not shown).

Using systemic murine infection experiments, we followed the persistence of bacteria on two different organs: the liver and the kidneys. After 7 days postinfection, 10 female mice were sacrificed, and their organs were removed, weighed, and homogenized in PBS buffer. Then, viable bacteria were enumerated. The numbers of CFU per gram of organ determined from the liver were not significantly different between the mutants and the wild type (Fig. 6A). In contrast, the number of Δef0685 or Δef1534 cells in the kidneys was lower than that of the wild-type strain. Indeed, a difference of 2 log₁₀ CFU per gram of organ for both mutants was observed (Fig. 6B). In conclusion, the involvement of these PPIases in colonization was tissue dependent, with the two foldases EF0685 and EF1534 having difficulties in colonizing the kidneys but not the liver. Involvement of EF0685 and EF1534 in kidney colonization could be explained by the folding of already identified surface-exposed virulence factors like Ace, the aggregation substance, or EfaA (6, 19, 34, 35) and/or other membrane proteins, including ABC transporters, which are necessary to acquire essential elements that could be in limited concentrations in vivo. A reduction of tissue-dependent colonization was already described for mutants affected in genes encoding SlrA or PpmA in S. pneumoniae. Indeed, both mutants showed a weakness to colonize mouse nasopharynx (7, 13). Interestingly, as observed here in E. faecalis, in S. pneumoniae, the PPIase which belongs to the cyclophilin family (SlrA) seemed to be less involved in colonization than the PPIase of the parvulin family (PpmA).

Fig 6.

Effect of ef0685 or ef1534 inactivation on the persistence of bacteria after infection. Mice were infected with the wild-type strain and Δef0685 and Δef1534 mutants. After 7 days of infection, the mice were sacrificed. Organs were removed and homogenized, and bacteria on agar plates were counted. The experiment was repeated 3 times. (A) Log₁₀ number of CFU per gram of liver; (B) log₁₀ number of CFU per gram of kidney.

To continue our investigation, we tested the cytotoxicity of the Δef0685 and Δef1534 mutants on Caco-2/T7 cells compared to that of the wild-type-strain. The eukaryotic cells and bacteria were coincubated for 24 h. The mortality of Caco-2/T7 cells was determined by measuring the amount of released LDH enzyme in the cell supernatant, and their viability was simultaneously quantified using a neutral red staining. The deficient ef0685 strain presented a clear loss of cytotoxicity in contrast to the wild-type strain (Fig. 7) and to the Δef1534 mutant (data not shown). Indeed, the neutral red staining showed a rate of 77.9% of eukaryotic cell survival when in contact with the Δef0685 mutant versus 9.5% when infected with the wild-type strain (Fig. 7A), while the LDH assay revealed 5.9% lysed eukaryotic cells after infection with the Δef0685 mutant versus 85.1% for the wild-type strain (Fig. 7B). EF0685 rotamase was clearly shown to be involved in cytotoxicity. This could be attributed to a direct involvement of the EF0685 PPIase in cytotoxicity or to its involvement in folding or in secretion of toxic secreted proteins. Therefore, in addition to common roles assigned to the extracellular PPIases, EF0685 is expected to be involved in specific functions contributing to cell cytotoxicity.

Fig 7.

Effect of ef0685 inactivation on cytotoxicity. Caco-2/T7 cells were infected with the E. faecalis wild type (black bars), Δef0685 mutant (white bars), and Δef0685C complemented strain (gray bars). (A) Evaluation of cell survival using neutral red staining; (B) evaluation of lysed cells using the reactive LDH method. Student's t test was used to determine whether results for the wild type and other strains were significantly different. *, P < 0.01.

Conclusion.

Results presented here underline the importance of the extracellular PPIases EF0685 and EF1534 for resistance to a high NaCl concentration and to ampicillin. In the mutant deficient in both the EF0685 and EF1534 PPIases, better growth under a high manganese concentration or under oxidative stress was observed, suggesting their involvement in conformation of transporters or oxidation targets.

The Δef0685 mutant displayed clear attenuated virulence traits in the G. mellonella model, difficulties in colonizing mouse kidneys, and loss of cytotoxicity, while the Δef1534 mutant showed weak attenuated virulence traits in G. mellonella and difficulties in colonizing mouse kidneys. Taken together, these data showed that EF0685 and EF1534 play a role in bacterial fitness and made E. faecalis surface PPIases interesting targets for therapeutic treatments.