Enterococcus faecalis rnjB Is Required for Pilin Gene Expression and Biofilm Formation

Strains, plasmids, growth media, and chemicals.The bacterial strains and plasmids used in this study are listed in Table 1. Brain heart infusion (BHI) broth and tryptic soy broth without glucose (TSB) were purchased from Difco Laboratories (Detroit, MI). All chemicals were purchased from Sigma (St. Louis, MO). Oligonucleotides used in this study were purchased from Invitrogen (La Jolla, CA) and are listed in Table 2.

Development of anti-EbpC MAbs.Recombinant EbpC (rEbpC) was produced in Escherichia coli as previously described by Nallapareddy et al. (35). To generate antibodies, BALB/c mice were immunized with the rEbpC protein using standard techniques (21). The splenocytes were collected and fused with SP2/O mouse myeloma cells as previously described (21). Monoclonal antibodies from single-cell clones were evaluated for binding specificity to rEbpC and to native antigen via enzyme-linked immunosorbent assays (ELISAs) and whole-cell ELISAs, respectively (19, 37) followed by kinetic screening for highest-affinity clones using surface plasmon resonance (SPR) as previously described (11).

Flow cytometry analysis.Aliquots equivalent to an optical density at 600 nm (OD₆₀₀) of 0.2 of E. faecalis cells in BHI medium were harvested and washed twice with phosphate-buffered saline (PBS) before resuspension in 100 μl of 1% bovine serum albumin (BSA) in PBS with 5 μg/ml anti-EbpC MAb 69 at 25°C for 1 h. This was followed by secondary labeling using a phycoerythrin-conjugated goat anti-mouse IgG (Jackson Immunoresearch, PA). The bacterial cells were fixed with 1% paraformaldehyde and analyzed with a BD FACSCalibur flow cytometer (BD Biosciences, CA).

Whole-cell ELISA library screen.The Tn insertion library in a 96-well format (17) kindly provided by D. A. Garsin was inoculated into 96-well U-bottom microplates (Greiner Bio-one, NC) containing 150 μl of BHI broth per well. Bacteria were cultured for 24 h before centrifugation and subsequently washed in 200 μl of PBS. Cells were then resuspended in 100 μl of 50 mM carbonate buffer, pH 9.6, and were used to coat Microlon 600 ELISA microtiter plates (Greiner Bio-one, NC) at 37°C for 2 h. Cells were washed 3 times with PBS-0.2% Tween 20 (PBST) and then incubated with 5% dry milk in PBST to block nonspecific binding. Anti-EbpC MAb 69 at 5 μg/ml in 5% dry milk in PBST was added to each well and incubated at 37°C for 1 h. The wells were washed 3 times with PBST, followed by the addition of 100 μl of a 1:3,000 dilution of goat anti-mouse IgG-horseradish peroxidase (HRP) (Jackson Immunoresearch). After 1 h of incubation at 37°C, the wells were washed 3 times with PBST. Tetramethylbenzidine (TMB) substrate was added, and the mixture was incubated at room temperature for 10 min. The absorbance of each well was measured at an OD₄₅₀.

Extraction of cell wall-associated proteins and Western blotting.Surface protein extracts from E. faecalis strains were prepared using mutanolysin as described previously with minor modifications (35). Briefly, the E. faecalis cells grown to stationary phase were harvested and washed with 0.02 M Tris-HCl (pH 7.0), 0.01 M MgSO₄ buffer and resuspended in 1/10 volume of this buffer containing 100 μM phenylmethylsulfonyl fluoride (PMSF). Following the addition of mutanolysin to a final concentration of 10 U/1 OD₆₀₀ of cells, tubes were incubated at 37°C for 1 h in a rotating shaker. After centrifugation at 12,000 × g for 10 min, the supernatants were concentrated by lyophilization, and aliquots were stored at −70°C. Equal amounts of total protein from mutanolysin extracts were loaded on 4% to 12% NuPAGE Novex bis-Tris gels (Invitrogen) under reducing conditions in MOPS (morpholinepropanesulfonic acid) buffer and transferred to an Immobilon-P polyvinylidene difluoride (PVDF) membrane (Millipore) according to the manufacturer's protocol. Membranes were then probed with anti-EbpC MAb followed by HRP-conjugated goat anti-mouse IgG antibody and developed using the TMB substrate system (KPL).

Construction of in-frame deletion mutants. E. faecalis OG1RF in-frame deletion mutants were obtained using the PheS counterselection markerless exchange system (22). First, ∼1-kb fragments of OG1RF chromosomal DNA flanking target genes were PCR amplified using the primer pairs listed in Table 2. The two amplicons for each gene were then simultaneously cloned into the counterselectable plasmid pCJK47 (22). The resulting plasmid was transformed into the conjugative donor strain CK111 (22) by electroporation. Blue colonies were used to conjugate with the recipient strain, OG1RF. Transconjugant colonies were selected for the integration of the plasmid and then cultured on MM9YEG agar supplemented with 10 mM p-chloro-phenylalanine. PCR analyses using primer pairs outside the targeted gene were performed on chromosomal DNA to identify mutant versus wild-type sequence at the target locus. PCR products were sequenced to confirm gene deletions.

Complementation analysis.The nisin-inducible E. faecalis expression vector pMSP3535 was applied in the complementation analysis according to the protocol of Bryan et al. (8). Briefly, a 1,699-bp fragment containing the rnjB gene and its ribosome binding site was PCR amplified from E. faecalis OG1RF and cloned under the control of the nisin promoter of the shuttle vector pMSP3535. Both the empty vector and the resulting plasmid were transformed into the rnjB deletion mutant. Nisin was added to the culture media at 25 ng/ml to induce exogenous expression of the rnjB gene. The presence of EbpC on the cell surface after nisin induction was determined by flow cytometry.

Northern blotting.An amount of cells grown in BHI to exponential phase that was equivalent to an OD₆₀₀ of ∼1.0 was collected for RNA extraction. Total RNA samples were extracted and blotted using standard methods and a NucleoSpin RNA II kit (Machery-Nagel GmbH & Co., Duren, Germany). Eight micrograms of RNA was analyzed by Northern blotting with ³²P-radiolabeled oligonucleotides specific for the 5′ untranslated region (5′-UTR) of ebpA (CAGTTAAATTAGAATTGCCTAGCACG), ebpC (GTCGTCGGTATGACCGTTATCA), or 23S rRNA (CGCCCTATTCAGACTCGCTTT). RNA levels were quantified using a Storm phosphorimager (GE Health Care), and the approximate sizes of the RNAs were determined by comparing their migration distances with those of the two rRNAs.

qRT-PCR.An amount of cells grown in BHI to exponential phase that was equivalent to an OD₆₀₀ of ∼1.0 was collected for RNA extraction. Total RNA and cDNA were prepared using methods provided by the manufacturer. Quantitative real-time PCR (qRT-PCR) on cDNA was performed using a SYBR green PCR master mix kit and an ABI7900 real-time PCR system (Applied Biosystems). The expression of 23S rRNA, ebpA, ebpB, and ebpC was analyzed using primer pairs listed in Table 2. For each primer set, a reference curve was established using the genomic DNA purified from wild-type OG1RF cells. The amounts of gene transcripts (ng/ml) obtained for ebpA, ebpB, and ebpC were normalized with the amount of 23S rRNA transcript.

β-Galactosidase assay.Cells containing the pTCV-lac vector or transcriptional fusions were grown overnight in BHI broth with kanamycin (1,000 μg/ml). Each was inoculated into BHI broth with kanamycin at a starting OD₆₀₀ of 0.05 and was harvested at mid-log phase and early stationary phase. For each strain, 10⁹ cells were washed and resuspended in 0.5 ml Z buffer (60 mM Na₂HPO₄, 40 mM NaHPO₄, 10 mM KCl, 1 mM MgSO₄, pH 7.0), mixed with a 0.25-ml volume of 0.1-mm-diameter zirconia beads (BioSpec Products, Bartlesville, OK), and disrupted using a minibeadbeater for 1 min. The cell lysates were centrifuged at 13,600 rpm for 1 min to remove cell debris before being used in a β-galactosidase assay. The β-galactosidase assay was performed according to the protocol of Bourgogne et al. (6), with some modifications. Dilutions of cell lysate were incubated with o-nitrophenyl-β-d-galactoside (ONPG) until a yellow color developed. The reaction was stopped by the addition of 0.5 volume of 1 M Na₂CO₃, and the absorbance of each well was measured at an OD₄₂₀. The total protein content of each cell lysate was determined using a standard bicinchoninic acid (BCA) assay to normalize samples as described previously (31). The β-galactosidase units of the samples corresponded to an OD₄₂₀/protein concentration in mg/ml.

Biofilm formation assay.A biofilm density detection assay was carried out as described previously by Mohamed et al. (32) with some modifications. Briefly, E. faecalis strains were grown overnight at 37°C before being diluted 1:100 into 10 ml TSBG (TSB plus 2% glucose) and cultured at 37°C in a Microtest 96-well tissue culture plate (BD Biosciences) for 24 h. The supernatant culture was carefully removed by gentle aspiration. The plate was washed four times with 200 μl PBS. The adherent cells were fixed with Bouin fixative for 30 min, washed two times with PBS, and then stained with 1% crystal violet for 30 min. The stain was rinsed and solubilized in ethanol-acetone (80:20). The absorbance of each well was measured at an OD₅₉₅. Each assay was performed in quadruplicate in three independent experiments.

Detection of surface EbpC through the use of an anti-EbpC monoclonal antibody.Using an anti-EbpC MAb, we were able to detect the presence of EbpC on the cell surfaces of most E. faecalis OG1RF cells grown under standard laboratory conditions in all growth phases (Fig. 1). The specificity and affinity of MAb 69 coupled with flow cytometry analysis provided a sensitive method to quantitatively evaluate the presence of EbpC on cell surfaces, and our results demonstrate that a higher percentage of cells express pili than previously reported using microscopy and polyclonal antibody labeling (35). Mean fluorescence intensity (MFI) in each growth phase was compared to that of an ebpABC deletion mutant control (Fig. 1). The result demonstrated that the surface display of EbpC is not growth phase related under these in vitro growth conditions. This observation provides evidence that EbpC is readily displayed in the E. faecalis population, which allowed us to consider the feasibility of a screen strategy to identify mutants deficient in surface pili.

• Open in new tab
• Download powerpoint

FIG. 1.

Cell surface binding of anti-EbpC MAb 69 to E. faecalis. Wild-type OG1RF (wt) and ebp deletion mutant (ΔebpABC) cells grown in BHI medium to early exponential phase, exponential phase, early stationary phase, and stationary phase were labeled with anti-EbpC MAb 69 and a secondary antibody-phycoerythrin conjugate, and 10,000 cells were analyzed by flow cytometry. Wild-type OG1RF cells grown to stationary phase and labeled with secondary antibody only (wild-type control) served as a negative control.

Identification of EbpC-deficient E. faecalis mutant strains.In an effort to identify the genetic factors that contribute to E. faecalis pilus gene expression and/or pilus assembly, we developed a whole-cell ELISA-based screening method using the anti-EbpC MAb 69. The 540-member Tn insertion mutant library used in the screen contains E. faecalis mutants that cover approximately one-fourth of all nonessential genes (17). Anti-EbpC MAb 69 was used to label cells from the insertion mutant library. Mutants which displayed the lowest levels of anti-EbpC MAb binding were selected. Figure 2 depicts the distribution of the ELISA signals of library members. Four mutants which repeatedly generated the lowest ELISA signals included those with insertions in ebpR (also known by its systematic name ef1090), ebpA (ef1091), ef1184, and ef1579 (Fig. 2). ebpR is a positive transcriptional regulator for the E. faecalis ebpABC operon (5), and thus its disruption would have a negative effect on EbpC expression. ebpA is the first gene of the ebpABC operon (35), and its disruption would have a polar effect on downstream genes, including ebpC. Of greatest interest was the identification of ef1184 and ef1579, representing E. faecalis dapA, which encodes a dihydrodipicolinate synthase, and lexA, which encodes a LexA repressor enzyme. Neither gene has been previously associated with pilus gene expression or pilus assembly in E. faecalis. In this study, we focus on the ef1184 insertion mutant.

• Open in new tab
• Download powerpoint

FIG. 2.

Identification of genes involved in the surface expression or regulation of EbpC. Each E. faecalis gene has a systematic name beginning with “ef,” and the 540 clones in the Tn library are represented by the number in their name on the x axis. Relative whole-cell ELISA signals are represented on the y axis. Each clone was labeled with our anti-EbpC antibody followed by a goat anti-mouse IgG-HRP conjugate and developed with the TMB substrate. Tn insertions in ef1090 (ebpR), ef1091 (ebpA), ef1184, and ef1579 (as indicated by arrows) were identified in two independent evaluations as clones which produced the lowest ELISA signal.

To confirm the whole-cell ELISA results, we analyzed each identified mutant strain using flow cytometry analysis with anti-EbpC MAb 69 staining (Fig. 3 A to E). The ef1184 insertion mutant showed an MFI that was 27% that of wild-type OG1RF. In comparison, ebpR and ebpA insertion mutants had MFIs of 3 and 5% of that for OG1RF, and a complete deletion of the ebpABC genes (35) resulted in an MFI of 0.3% of that of the wild type. Thus, the level of surface-exposed EbpC is reduced in the three insertion mutants. The decrease in surface-displayed EbpC in these insertion mutants was further confirmed by Western blotting (Fig. 4). Ladders of high-molecular-weight covalently linked EbpC were observed in cell wall extract fractions of wild-type OG1RF and to a lesser extent in the ef1184 insertion mutant, reflecting the 3- to 4-fold decrease seen in fluorescence intensity by flow cytometry of the ef1184 mutant population compared to that of the wild type. Surface extracts of both the ebpA and ebpR insertion mutants exhibited a single low-molecular-weight band likely representing the EbpC monomer (molecular mass, 65 kDa), indicating the deficiency of pilus polymer formation in these two mutants. All together, ELISA, flow cytometry, and Western analyses indicated that the insertion in ef1184 caused a reduction in the level of surface-exposed EbpC.

• Open in new tab
• Download powerpoint

FIG. 3.

Flow cytometry analysis of surface-exposed EbpC in E. faecalis wild-type and mutant strains. Each population was labeled with anti-EbpC MAb 69. Mean fluorescence intensities are compared in the OG1RF (A), ΔebpABC (ebpABC deletion) (B), ebpR (ef1090 insertion mutant) (C), ebpA (ef1091 insertion mutant) (D), ef1184 (dapA insertion mutant) (E), ΔdapA (dapA deletion mutant) (F), and ΔrnjB (rnjB deletion mutant) (G) strains. The MFI value of each sample is indicated. ▾, Tn insertion mutant.

• Open in new tab
• Download powerpoint

FIG. 4.

Western blotting results of Ebp pilus formation of the wild-type OG1RF and mutant strains. Mutanolysin surface extractions were probed with anti-EbpC MAb 69. Lanes (left to right): EZ-run protein marker (M), OG1RF, ebpR mutant, ebpA mutant, ef1184 mutant, ΔebpABC mutant, ΔrnjB mutant, and ΔdapA mutant. High-molecular-weight Ebp pilus polymers and EbpC monomers are indicated. ▾, Tn insertion mutant.

Deletion of rnjB reduces the level of surface-exposed EbpC.Sequence analysis of the 3.7-kb gene locus containing ef1184 predicted the presence of three complete open reading frames, asd (ef1183), dapA (ef1184), and rnjB (ef1185). No transcriptional terminator-like sequence can be identified in between the open reading frames, suggesting that they are transcribed together as one polycistronic mRNA (Fig. 5 A). This gene cluster organization is similar to that of a B. subtilis diaminopimelate operon which contains four genes, asd, dapG, dapA, and rnjB (previously ymfA) (Fig. 5A) (12). Both asd (encodes an aspartate-semialdehyde dehydrogenase) and dapA (encodes a dihydrodipicolinate synthase) are involved in the synthesis of the peptidoglycan precursor diaminopimelate, while the gene product of rnjB in B. subtilis was identified as a novel endoribonuclease, RNase J2 (16). BLASTN search using the sequence of the OG1RF ef1183-ef1185 locus revealed >99% identity among all 23 known E. faecalis genomes, which clearly demonstrates the ubiquitous presence of this cluster among E. faecalis strains.

• Open in new tab
• Download powerpoint

FIG. 5.

E. faecalis rnjB encodes an RNase J2 ortholog. (A) Comparison of the E. faecalis OG1RF ef1183-ef1185 operon with the B. subtilis asd-rnjB operon. The red arrow indicates the Tn insertion site in the ef1184 mutant (dapA insertion mutant). (B) Phylogram of six RNase J1s and six RNase J2s from the bacterial order Bacilli and several other bacterial RNase Js based on multiple-sequence alignment generated by ClustalW2 with default parameters.

The Tn917 insertion site in the dapA (ef1184) insertion mutant was located at +169 bp in the dapA gene (17). A polar effect of the Tn insertion in the dapA gene may also alter the expression of the downstream rnjB gene. To determine which gene is responsible for the reduced EbpC surface display seen in the insertion mutant, we constructed two in-frame deletion mutants with either dapA (ΔdapA mutant) or rnjB (ΔrnjB mutant) from the wild-type OG1RF strain deleted. Both mutants are viable, indicating that dapA and rnjB are not essential genes for E. faecalis. The EbpC surface display of each mutant was analyzed by flow cytometry. Interestingly, the ΔrnjB mutant exhibited a low level of EbpC on the cell surface (5% of the wild-type MFI) (Fig. 3G). In contrast, the ΔdapA mutant showed a level of anti-EbpC labeling similar to that of wild-type OG1RF (Fig. 3F). To confirm the importance of the rnjB gene in altering levels of surface-exposed EbpC, the gene was cloned into the nisin-inducible expression vector pMSP3535 and introduced into the ΔrnjB mutant. As shown in Fig. 6, the in trans addition of rnjB in the ΔrnjB mutant partially restored levels of surface-exposed EbpC (Fig. 6D and E).

• Open in new tab
• Download powerpoint

FIG. 6.

Complementation of the rnjB deletion in OG1RF. Flow cytometry analysis of OG1RF (wild type) (B), the ΔrnjB mutant (C), the ΔrnjB mutant with the vector control (D), and the ΔrnjB mutant with the rnjB complement vector (E). Cells grown to the stationary phase in BHI medium with 25 ng/ml nisin were labeled with anti-EbpC MAb 69 followed by a secondary antibody-phycoerythrin conjugate. Mean fluorescence intensities are compared to that of the control, OG1RF with secondary antibody only (A).

The effect of rnjB on pili was further confirmed by Western blotting. As shown in Fig. 4, ladders of high-molecular-weight bands representing the Ebp pilus polymers were detected in the cell wall extractions of both wild-type OG1RF and the ΔdapA mutant probed with anti-EbpC MAb, which demonstrated that pilus formation is not affected by the deletion of dapA. In comparison, these high-molecular-weight bands are absent in cell wall extracts of the ΔrnjB mutant, similar to what occurs in the ebpABC deletion mutant (Fig. 4). This indicates that the formation of the EbpC-containing pilus is attenuated in the ΔrnjB mutant. Considering all data together, we conclude that rnjB somehow regulates the level of surface-exposed EbpC and that the loss of EbpC on the cell surface of the dapA Tn insertion mutant is the result of its polar effect on the downstream rnjB gene.

As an initial step toward characterizing the function of RnjB, we carried out sequence comparisons using BLASTP and ClustalW. Multiple sequence alignments (not shown) clearly showed that RnjB belongs to the RNase J family. Many Gram-positive bacteria, including Bacillus subtilis, Staphylococcus aureus, Streptococcus pneumoniae, Streptococcus pyogenes, E. faecalis, and Listeria monocytogenes, contain two RNase J family members. E. faecalis RnjB is a member of the RNase J2 cluster (Fig. 5B). The orthology of E. faecalis RnjB and B. subtilis RnjB is also supported by their presence in similar operons.

rnjB affects ebpABC gene expression.The effect of rnjB on surface-displayed EbpC may occur at the RNA level and affect processes such as transcriptional regulation and RNA processing, or it may occur at the protein level and affect processes such as translation, transportation, and pilus assembly. To start to elucidate the mechanism of the rnjB effect on EbpC surface display, we explored the steady-state mRNA levels for ebp genes. Figure 7 A shows the results of a Northern blot analysis of RNAs isolated from wild-type OG1RF and the ΔrnjB mutant. Both an ebpA 5′-UTR probe and an ebpC probe detect an ∼7-kb mRNA in the wild-type OG1RF RNA blot, which presumably represents the polycistronic ebpABC mRNA (Fig. 7A). In comparison, the intensity of that band is strongly reduced in the ΔrnjB mutant RNA blot probed with either probe (Fig. 7A). These results clearly demonstrate that the ebpABC transcript level is greatly decreased in the ΔrnjB mutant compared to that of wild-type OG1RF. qRT-PCR analysis further demonstrated that the steady-state mRNA levels of ebpA, ebpB, and ebpC were, respectively, 36-, 24-, and 16-fold higher in OG1RF than in the ΔrnjB mutant (Fig. 7B). Thus, the expression of all three ebp pilin genes is affected by rnjB at the mRNA level.

• Open in new tab
• Download powerpoint

FIG. 7.

Comparison of transcript levels of ebp-related genes in wild-type OG1RF and the ΔrnjB mutant. (A) Northern blot of RNA isolated from mid-log-phase wild-type OG1RF and ΔrnjB cells. Blots were hybridized with ³²P-labeled 5′-UTR-ebpA (top panel), ebpC (middle panel), and 23S rRNA (loading control) (bottom panel) probes. (B) qRT-PCR analysis of RNA isolated from mid-log-phase wild-type OG1RF and ΔrnjB cells. Each column represents the gene of interest shown in the x axis. The fold decrease in gene expression in the ΔrnjB mutant corresponds to the ratio of the transcript level of each gene in the wild-type strain to that of each gene in the ΔrnjB mutant. The level of each gene transcript is tested in triplicate and normalized using 23S rRNA transcript levels.

We next evaluated the effect of rnjB on an ebpA::lacZ transcriptional fusion. pTEX5585 is a pTCV-lac-derived vector containing the ebpA promoter, the 5′-UTR, and the first 41 nucleotides (nt) of the coding region upstream of a LacZ coding region (6). As shown in Fig. 8 A, LacZ expression from this reporter in OG1RF is 20-fold higher than in the ΔrnjB mutant at both log phase and stationary phase. Since ebpA is known to be regulated by EbpR, we also analyzed a similar ebpR::lacZ transcriptional fusion. Figure 8 shows that ebpR::lacZ expression is not affected by ΔrnjB and thus that the effect on ebpA::lacZ is specific. Furthermore, this result shows that the effect of rnjB on ebpABC expression is not an indirect effect of higher ebpR transcription.

• Open in new tab
• Download powerpoint

FIG. 8.

rnjB affects the expression of an ebpA::lacZ reporter but not an ebpR::lacZ reporter. Wild-type OG1RF and the ΔrnjB mutant containing either P_ebpR::lacZ or P_ebpA::lacZ were grown in BHI and collected at log phase and stationary phase for β-galactosidase (B-gal) assay. The y axis represents the β-galactosidase units (OD₄₂₀/total protein concentration in mg/ml). Error bars represent the standard errors of the measurements from three individual cultures.

rnjB is essential for E. faecalis biofilm formation.To evaluate the functional impact of the rnjB deletion on virulence-related phenotypes, we evaluated the in-frame deletion mutant to monitor differences in the abilities of the strains to form biofilm on a polystyrene surface. Utilizing a polystyrene biofilm assay, it has been previously demonstrated that E. faecalis Ebp pili are required for biofilm formation (35). Figure 9 demonstrates that, compared to the wild type, the ΔrnjB mutant is attenuated in biofilm formation, a result similar to that of the ebpABC deletion mutant. In contrast, the dapA deletion mutant demonstrated biofilm formation similar to that of the wild type, consistent with our findings that the loss of the pilus phenotype of the dapA Tn insertion mutant is a result of its polar effect on rnjB expression.

• Open in new tab
• Download powerpoint

FIG. 9.

Biofilm formation of OG1RF gene deletion mutants. Cells grown for 24 h on polystyrene microtiter plates were evaluated for biofilm formation, expressed as a mean crystal violet absorbance of a biofilm biomass in four independent microtiter wells. Error bars represent the standard errors of the means.