Bacteriocin Protein BacL₁ of Enterococcus faecalis Targets Cell Division Loci and Specifically Recognizes l-Ala₂-Cross-Bridged Peptidoglycan

Abstract

Bacteriocin 41 (Bac41) is produced from clinical isolates of Enterococcus faecalis and consists of two extracellular proteins, BacL₁ and BacA. We previously reported that BacL₁ protein (595 amino acids, 64.5 kDa) is a bacteriolytic peptidoglycan d-isoglutamyl-l-lysine endopeptidase that induces cell lysis of E. faecalis when an accessory factor, BacA, is copresent. However, the target of BacL₁ remains unknown. In this study, we investigated the targeting specificity of BacL₁. Fluorescence microscopy analysis using fluorescent dye-conjugated recombinant protein demonstrated that BacL₁ specifically localized at the cell division-associated site, including the equatorial ring, division septum, and nascent cell wall, on the cell surface of target E. faecalis cells. This specific targeting was dependent on the triple repeat of the SH3 domain located in the region from amino acid 329 to 590 of BacL₁. Repression of cell growth due to the stationary state of the growth phase or to treatment with bacteriostatic antibiotics rescued bacteria from the bacteriolytic activity of BacL₁ and BacA. The static growth state also abolished the binding and targeting of BacL₁ to the cell division-associated site. Furthermore, the targeting of BacL₁ was detectable among Gram-positive bacteria with an l-Ala-l-Ala-cross-bridging peptidoglycan, including E. faecalis, Streptococcus pyogenes, or Streptococcus pneumoniae, but not among bacteria with alternate peptidoglycan structures, such as Enterococcus faecium, Enterococcus hirae, Staphylococcus aureus, or Listeria monocytogenes. These data suggest that BacL₁ specifically targets the l-Ala-l-Ala-cross-bridged peptidoglycan and potentially lyses the E. faecalis cells during cell division.

INTRODUCTION

Enterococcus faecalis is a commensal Gram-positive bacterium in the intestinal tract of healthy humans or animals and is also known to be an opportunistic pathogen causing various infectious diseases, including urinary infectious disease, bacteremia, infective endocarditis, and others (1–3). The infection-derived E. faecalis strains often produce various plasmid-encoded bacteriocins (4, 5).

Bacteriocins are bacterial peptides or proteins with antimicrobial activities (6). Heat- and acid-stable bacteriocin peptides produced by Gram-positive bacteria are divided into class I and class II according to posttranslational modifications (7, 8). Class I bacteriocins are lantibiotics that contain nonproteinogenic amino acids generated by posttranslational modification (9). Only two class I bacteriocins have been identified in enterococci: β-hemolysin/bacteriocin (cytolysin) and enterocin W (10–14). In contrast, most enterococcal bacteriocins belong to class II and are nonmodified antimicrobial peptides, such as AS-48, enterocin A, and others (7, 15, 16). We have found the enterococcal class II bacteriocins, including Bac21, Bac31, Bac32, Bac43, and Bac51, in clinical strains of E. faecalis or Enterococcus faecium (17–21). Unlike the low-molecular-weight peptide-type class I and II bacteriocins, heat-labile antimicrobial proteins are referred to as bacteriolysins, previously named class III bacteriocins, and show enzymatic bactericidal activity (22, 23). In enterococci, the bacteriolysins enterolysin A and bacteriocin 41 (Bac41) have been identified (24–26).

Bac41 was originally found expressed from the pheromone-responsive plasmid pYI14 carried by the clinical strain E. faecalis YI14 (26, 27). The Bac41-type bacteriocins were also found in the E. faecalis VanB-type vancomycin-resistant E. faecalis (VRE) outbreak strains (27). Bac41 is specifically active only against E. faecalis (26, 28). The determinant region of Bac41 contains six open reading frames (ORFs), including bacL₁, bacL₂, bacA, and bacI (Fig. 1A). The bactericidal activity of Bac41 is actually expressed by the two extracellular components, the bacL₁- and bacA-encoded proteins BacL₁ and BacA (26). BacL₁ and BacA are secreted proteins that coordinately exert bactericidal activity against E. faecalis (26, 28). BacL₂ positively regulates the transcripts of bacL₁ and bacL₂ itself (unpublished data). BacI is the specific immunity factor protecting a Bac41 producer from Bac41 activity (26).

FIG 1.

Schematics of Bac41 gene organization and BacL₁ structure. (A) Organization of Bac41-related genes (GI 169635857). (B) Molecular structure of BacL₁ (GI 169635864). Two domains with homology to distinct peptidoglycan hydrolases, bacteriophage-type hydrolase and NlpC/P60 family hydrolase, are present in the regions from amino acid (a.a.) 3 to 140 and amino acid 163 to 315, respectively. Three repeats of the bacterial SH3 domain are present in the region from amino acid 329 to 590.

We previously demonstrated that BacL₁ is a peptidoglycan d-isoglutamyl-l-lysine endopeptidase (28). BacL₁ has 595 amino acids with a molecular mass of 64.5 kDa and consists of two distinct peptidoglycan hydrolase homology domains and three repeats of the SH3 domain (Fig. 1B). The two peptidoglycan hydrolase domains located in the regions from amino acid 3 to 140 and amino acid 163 to 315 show homology to the bacteriophage-type peptidoglycan hydrolase and the NlpC/P60 family peptidoglycan hydrolase, respectively (26, 29, 30). The second peptidoglycan hydrolase homologue, with similarity to NlpC/P60, has d-isoglutamyl-l-lysine endopeptidase activity against the purified peptidoglycan component from E. faecalis (28). On the other hand, the molecular function of the first peptidoglycan hydrolase domain, with similarity to bacteriophage-type peptidoglycan hydrolase, remains to be elucidated but is still required for the bactericidal activity against viable E. faecalis cells (28). The SH3 repeat domain is located in the region from amino acid 329 to 590 and functions as the binding domain to the peptidoglycan (28). However, BacL₁ is not sufficient for bactericidal activity. BacA appears to be essential for bactericidal activity, together with BacL₁, although its function also remains to be determined (26, 28).

On the basis of cell morphology, enterococci are grouped in the ovococci, whose cell shapes are elongated ellipsoids (31–33). In ovococci, the model of the dividing cell wall assembly process is distinct from that of other shaped bacteria, such as spherical cocci. The cell division of ovococci is achieved by two distinct cell wall-synthesizing machineries that manage peripheral and septal cell wall growth. The peripheral cell wall growth is responsible for the longitudinal cell elongation. On the other hand, the septal cell wall growth occurs to allow splitting into separated daughter cells. In this study, by using fluorescent dye-conjugated recombinant proteins, we demonstrated that BacL₁ localized to the cell division-related cell surface of target E. faecalis cells and that cell division was required for susceptibility to the bactericidal activity expressed by BacL₁ and BacA.

MATERIALS AND METHODS

Bacterial strains, plasmids, oligonucleotides, media, and antimicrobial reagents.

The bacterial strains and plasmids used in this study are shown in Table 1. A standard plasmid DNA methodology was used (34). Enterococcal strains were routinely grown in Todd-Hewitt broth (THB; Difco, Detroit, MI) at 37°C (35), unless otherwise noted. Escherichia coli strains were grown in Luria-Bertani medium (LB; Difco) at 37°C. Gram-positive bacterial species (other than Enterococcus) were grown in brain heart infusion (BHI) medium (Difco) at 37°C. The antibiotic concentrations for the selection of E. coli were 100 mg liter⁻¹ ampicillin and 30 mg liter⁻¹ chloramphenicol. The concentration of chloramphenicol for the routine selection of E. faecalis carrying plasmid pAM401 or its derivatives was 20 mg liter⁻¹, unless otherwise noted. All antibiotics were obtained from Sigma Co. (St. Louis, MO).

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Description	Source or reference
Strains
E. faecalis OG1S	str, derivative of OG1	35
E. faecalis OG1X	str, protease-negative derivative of OG1	35
E. faecalis OG1RF	rif fus, derivative of OG1	35
E. faecalis FA2-2	rif fus, derivative of JH2	60
E. faecium BM4105RF	rif fus, derivative of BM4105	61
E. hirae 9790	Type strain	ATCC 9790
S. aureus F-182	Clinical isolate, resistant to methicillin and oxacillin	ATCC 43300
S. pyogenes MGAS315	Clinical isolate, serotype M3	ATCC BAA-595
S. pneumoniae 262	Quality control strain, serotype 19F	ATCC 49619
L. monocytogenes EGD	Serovar 1/2a	ATCC BAA-679
E. coli DH5α	endA1 recA1 gyrA96 thi-1 hsdR17 supE44 relA1 Δ(argE-lacZYA)U169, host for DNA cloning	Bethesda Research Laboratories
E. coli BL21	ompT hsdSB(rB− mB−) gal(λcI 857 ind1 Sam7 nin5 lacUV5-T7gene1) dcm(DE3), host for protein expression	Novagen
Plasmids
pAM401	E. coli-E. faecalis shuttle plasmid; cat tet	62
pHT1100	pAM401 derivative containing wild-type Bac41 genes	26
pET22b(+)	Expression plasmid for His-tagged protein in E. coli	Novagen
pET::bacL1	pET22b(+) derivative expressing BacL1	28
pET::bacL1 Δ1	pET22b(+) derivative expressing BacL1Δ1	28
pET::bacL1 Δ2	pET22b(+) derivative expressing BacL1Δ2	28
pET::bacL1 Δ1Δ2	pET22b(+) derivative expressing BacL1Δ1Δ2	28
pET::bacL1 Δ3	pET22b(+) derivative expressing BacL1Δ3	28
pET::bacA	pET22b(+) derivative expressing BacA	28

Recombinant proteins and antibodies.

The histidine-tagged recombinant proteins of full-length BacL₁, its truncated derivatives, and BacA were prepared by the Ni-nitrilotriacetic acid (NTA) system as previously described (28). The green or red fluorescent dye-labeled recombinant proteins were prepared with NH₂-reactive fluorescein or NH₂-reactive HiLyte Fluor 555 (Dojindo, Kumamoto, Japan), respectively. By performing a soft-agar bacteriocin assay, we confirmed that fluorescent dye-conjugated BacL₁ remains active (see Fig. S1 in the supplemental material). Anti-BacL₁ antibody was prepared by immunization of rabbits with recombinant BacL₁-His protein as previously described (Operon Technologies, Alameda, CA) (28).

Fluorescence microscopy.

Bacteria diluted with fresh medium were mixed with fluorescent recombinant protein as indicated and incubated at 37°C for 1 h. The bacteria were collected by centrifugation at 5,800 × g for 3 min and then fixed with 4% paraformaldehyde at room temperature (RT) for 15 min. The bacteria were rinsed and resuspended with distilled water and mounted with Prolong gold antifade reagent with 4′,6-diamidino-2-phenylindole (DAPI; Invitrogen, Carlsbad, CA) on a glass slide. The sample was analyzed by fluorescence microscopy (Axiovert 200; Carl Zeiss, Oberkochen, Germany), and images were obtained with a DP71 camera (Olympus, Tokyo, Japan).

Immunogold TEM.

Bacteria in early exponential phase were treated with recombinant BacL₁ and BacA as indicated and incubated at 37°C for 1 h. The bacteria were fixed with 3% paraformaldehyde–0.1% glutaraldehyde for 10 min at RT and mounted on an electron microscopy (EM) grid. After fixation, the sample grid was treated with 10-fold-diluted anti-BacL₁ antibodies in phosphate-buffered saline (PBS) containing 2% bovine serum albumin (BSA) at 37°C for 1 h, followed by a wash with PBS. Then, the grid was treated with 10-fold diluted colloidal gold (15 nm)-conjugated anti-rabbit IgG in PBS containing 2% BSA for 30 min at RT, washed with PBS, and then negatively stained with 2% ammonium molybdate for 1 min at RT. The resulting samples were analyzed by transmission electron microscopy (TEM) (JEM-1010; JEOL Ltd., Tokyo, Japan).

Bacteriolytic assay.

The soft-agar assay or liquid-phase assay for bacteriocin activity was performed as described previously (36). Briefly, the test bacterial strain or 1 μl of the recombinant protein solution was inoculated onto THB soft agar (0.75%) containing the indicator strain and was then incubated at 37°C for 24 h. The formation of an inhibitory zone was evaluated as susceptibility to the bacteriocin. For the agar-based bacteriolytic assays using Streptococcus pyogenes and Streptococcus pneumoniae, the indicator bacteria were spread on agar plates by swab instead of using the soft agar. In this swab method, E. faecalis OG1S and E. faecium BM4105 RF were used for the positive control and the negative control, respectively. For the liquid-phase bactericidal assay, an overnight culture of the indicator strain was diluted with fresh medium, and then the recombinant proteins were added and the sample was incubated at 37°C. Changes in turbidity were monitored by using a spectrometer (DU730; Beckman Coulter, Fullerton, CA) or microplate reader (Thermo Scientific, Waltham, MA).

Cell wall degradation assay.

For the cell wall degradation assay, a cell wall fraction was prepared as described previously, with slight modifications (28, 37). The bacterial culture was collected by centrifugation and rinsed with 1 M NaCl. The bacterial pellet was suspended in 4% SDS and heated at 95°C for 30 min. After rinsing with distilled water four times, the bacterial pellet was resuspended with distilled water and mechanically disrupted with 0.1-mm glass beads (As One, Osaka, Japan) using a FastPrep FP100A (Thermo Scientific, Waltham, MA). After unbroken cells were removed by centrifugation at 1,000 rpm for 1 min, the cell wall fraction in the supernatant was collected by centrifugation at 15,000 rpm for 10 min and was then treated with 0.5 mg ml⁻¹ trypsin (0.1 M Tris-HCl [pH 6.8], 20 mM CaCl₂) at 37°C for 16 h. The sample was further washed with distilled water four times and was resuspended in 10% trichloroacetic acid (TCA), followed by incubation at 4°C for 5 h, and then given additional washes with distilled water four times (38). Finally, the cell wall fraction was resuspended in PBS and quantified by measuring the turbidity for the cell wall degradation assay. Mutanolysin (Sigma) was used as a positive control for the cell wall degradation enzyme.

RESULTS

BacL₁ targets the cell division-associated region on the E. faecalis surface via its cell wall binding domain.

To investigate the localization of BacL₁ on target E. faecalis cells, we coincubated E. faecalis cells and the recombinant BacL₁ labeled with red fluorescent dye in the absence or presence of BacA, followed by analysis using fluorescence microscopy (Fig. 2A and B). A specific localization signal of BacL₁ in the midcell was observed independently of BacA (Fig. 2A). Furthermore, the four characteristic localization patterns closely correlated with cell growth division were detected (31, 33, 39). First, the most typical localization signal of BacL₁ was detected in the midcell, which corresponds to the equatorial ring (Fig. 2B). The equatorial ring structure of the BacL₁ localization in the midcell was clearly recognized by the reconstructed image of fluorescence microscopy analysis (see Movie S1 in the supplemental material). Second, the duplicated equatorial ring structure was detected as the source of the localization signal of BacL₁ in the cells initiating elongation prior to cell division. Third, in the cells where the cell division process had progressed further, to formation of the division septum, the localization signal of BacL₁ was distributed in the area from the equatorial ring to the division septum, where the cell wall is newly synthesized (nascent cell wall) (32). Furthermore, when cell division was completed, localization at the division septum between separated daughter cells, as well as at the equatorial ring, was detected. In addition, immunogold TEM analysis using anti-BacL₁ antibodies in E. faecalis cultures treated with BacL₁ and BacA also showed the equatorial ring localization of BacL₁ (Fig. 2C).

FIG 2.

Localization of BacL₁ on the E. faecalis cell surface. (A) An overnight culture of E. faecalis OG1S diluted 5-fold with fresh THB broth was incubated with HiLyte Fluor 555 fluorescent dye-labeled (red) BacL₁ (5 μg/ml) in the presence (bottom) or absence (middle) of BacA, followed by analysis using fluorescence microscopy. Bacteria grown without red fluorescent conjugate are also shown as a negative control (top). Phase contrast (Ph) is pseudocolored (green) in the merged image. (B) Extensive representation of the localization pattern of red fluorescent dye-labeled BacL₁. The sample preparation was performed exactly as described for panel A. DNA was visualized with DAPI (blue). The schematic on the right illustrates the four characteristic patterns of BacL₁ localization and cell division states. (C) E. faecalis treated with BacL₁ and BacA (5 μg/ml each) was subjected to immunogold transmission electron microscopy using anti-BacL₁ antibodies. The arrow indicates the septum localization of gold particles.

We previously reported that BacL₁ binds to peptidoglycan of E. faecalis via a C-terminal SH3 triple repeat domain localized in the region from amino acid 329 to 590 (28). To investigate the domain required for the specific targeting, domain deletion derivatives of BacL₁ were labeled with green fluorescent dye (Fig. 3A) and mixed with E. faecalis cells, followed by analysis of their location signal by fluorescence microscopy (Fig. 3B). BacL₁Δ3, the derivative with deletion of the C-terminal SH3 repeat, failed to localize to the equatorial ring and did not show any detectable signal. In contrast, BacL₁Δ1, BacL₁Δ2, and BacL₁Δ1Δ2, derivatives with deletion of the phage-type peptidoglycan hydrolase homology domain, NlpC/P60 family peptidoglycan hydrolase homology domain, or both domains, respectively, were targeted to the equatorial ring similarly to wild-type BacL₁. These results indicate that the SH3 repeat was sufficient for the targeting to the equatorial ring on the cell surface of E. faecalis. Collectively, BacL₁ appeared to target the cell division-related cell surface, including the equatorial ring, the division septum, and the nascent cell wall, via its C-terminal SH3 repeat domains.

FIG 3.

Domain of BacL₁ that is responsible for septum targeting. (A) Schematics of truncated BacL₁ constructs. (B) Overnight culture of E. faecalis OG1S diluted 5-fold with fresh THB broth was incubated with the fluorescein dye-labeled (green) truncated BacL₁ proteins (5 μg/ml) depicted in panel A, followed by analysis using fluorescence microscopy. Phase contrast (Ph) is pseudocolored (red) in merged images.

Cell division is required for the septum targeting of BacL₁ and for the cell lysis triggered by BacL₁ and BacA.

To analyze the involvement of cell division in Bac41 activity, we investigated the relationship of growth phase and susceptibility to Bac41-induced lysis. E. faecalis was grown in fresh THB broth, and a mixture of recombinant BacL₁ and BacA was added at different points in the growth phases (Fig. 4A). Adding BacL₁ and BacA at the start of incubation (time zero) completely inhibited the increase of the bacterial suspension's turbidity (cell growth). When BacL₁ and BacA were added at early or mid-exponential phase, bacterial turbidity was also dramatically decreased, indicating that cells were lysed. In contrast, treatment with BacL₁ and BacA at the stationary phase did not affect the bacterial turbidity, similar to the results for the untreated culture. The bacterial viability test by colony formation assay also indicated that the bactericidal activity of BacL₁ and BacA was effective only in early or exponential phase but not stationary phase (Fig. 4B). In contrast, egg white lysozyme was able to decrease the viability of bacteria even in stationary phase (Fig. 4B). These observations indicated that E. faecalis in stationary phase was not susceptible to the cell lysis induced by BacL₁ and BacA. Then, to test the growth phase dependence of the septum targeting of BacL₁, the red fluorescence-labeled BacL₁ was incubated with E. faecalis cells in early exponential or stationary phase, and the BacL₁ localization was analyzed by fluorescence microscopy (Fig. 4C). In the case of the bacteria in early exponential phase, BacL₁ localized at the division septum. In contrast, the septum localization of BacL₁ was not observed in bacteria in stationary phase. BacL₁ also failed to even bind to the cell surface in stationary-phase bacteria (Fig. 4C). These results suggested that BacL₁ recognized the dividing cell surface. Furthermore, we investigated the susceptibility to bactericidal activity of BacL₁ and BacA when bacterial cell growth was artificially restricted with various antibiotic reagents (Fig. 5). Treatment with bacteriostatic antibiotics, such as chloramphenicol and tetracycline, almost completely rescued the cells from the bacteriolytic activity of BacL₁ and BacA (Fig. 5A and B). The localization of BacL₁ to the equatorial ring was also abolished in the chloramphenicol- or tetracycline-treated bacteria (Fig. 5C). Treatment with vancomycin, a bactericidal drug blocking cell wall synthesis, resulted in relief of the sensitivity of E. faecalis to lysis by BacL₁ and BacA and abolished BacL₁ targeting to the cell surface (Fig. 5A, B, and C). Interestingly, the bacteria treated with ampicillin, which has an elongating effect on bacterial cells by inhibiting the penicillin binding protein (PBP) functions, appeared to be more susceptible to the bactericidal activity of BacL₁ and BacA (Fig. 5A and B) and to the septum targeting of BacL₁ (Fig. 5C).

FIG 4.

Growth phase dependence of the susceptibility to Bac41. (A) An overnight culture of E. faecalis OG1S diluted 100-fold with fresh THB broth was incubated at 37°C. A mixture of recombinant BacL₁ and BacA (5 μg/ml each) was added at different growth phases corresponding to 0 h, 2 h, 3.5 h, or 5 h, as indicated with arrows. An untreated culture served as the negative control. The turbidity (optical density at 600 nm [OD₆₀₀]) was monitored in each culture. The data are presented as the mean results ± standard deviations (SD) of three independent experiments. (B) E. faecalis was treated with BacL₁ and BacA at different growth phases as described for panel A. After further incubation for 1 h from each time point of addition, the bacterial suspensions were serially diluted 10-fold with fresh THB broth and then spotted onto a THB agar plate, followed by incubation overnight. Colony formation was evaluated as a measure of bacterial viability. Lysozyme was used as a control. (C) E. faecalis was treated with HiLyte Fluor 555-labeled (red) BacL₁ (5 μg/ml) in the early-exponential (2 h) or stationary (5 h) phase of growth. After further incubation for 1 h from each time point of addition, the cells were fixed and analyzed by fluorescence microscopy. Phase contrast (Ph) is pseudocolored (green) in the merged images.

FIG 5.

Effects of antibiotics on the susceptibility to Bac41. (A) An overnight culture of E. faecalis OG1S diluted 5-fold with fresh THB broth was incubated with (Bac+) or without (Bac−) a mixture of recombinant BacL₁ and BacA (5 μg/ml each) in the presence or absence of ampicillin (ABPC; 20 μg/ml), chloramphenicol (CP; 100 μg/ml), tetracycline (TC; 12.5 μg/ml), or vancomycin (VCM; 10 μg/ml). The turbidity at 600 nm was measured with a microplate reader during the incubation period. The data are presented as the mean results ± SD of three independent experiments. (B) E. faecalis was treated with (+) or without (−) a mixture of BacL₁ and BacA in the presence of antibiotics as described for panel A. After incubation for 6 h, the bacterial suspensions were serially diluted 10-fold with fresh THB broth and then spotted onto a THB agar plate, followed by incubation overnight. Colony formation was evaluated as a measure of bacterial viability. (C) An overnight culture of E. faecalis diluted 5-fold with fresh THB broth was treated with HiLyte Fluor 555-labeled (red) BacL₁ (5 μg/ml) in the presence of antibiotics as shown. After incubation for 1 h, the cells were fixed and analyzed by fluorescence microscopy. Phase contrast (Ph) is pseudocolored (green) in merged images.

Specific recognition by BacL₁ of l-Ala₂-type peptidoglycan cross-bridging structure.

The composition and length of the cross-bridge peptide-linking stem peptides bound to N-acetyl-muramic acid are diverse among bacterial species (Fig. 6A) (40, 41). Lu et al. reported that the SH3 domain of ALE-1, a bacteriolytic peptidoglycan hydrolase of Staphylococcus aureus, specifically recognizes the pentaglycine cross bridge, which is a specific structure in the peptidoglycan of S. aureus (42). As shown by the results in Fig. 3, the SH3 domains appeared to be necessary for targeting the cell division-related region. To investigate whether the SH3 domain of BacL₁ also specifically recognizes the cross-bridging structure in the peptidoglycan of E. faecalis, we analyzed the cell division-associated targeting of BacL₁ in various Gram-positive bacterial species, including E. faecalis OG1S, E. faecalis OG1X, E. faecalis OG1RF, E. faecalis FA2-2, E. faecium BM4105RF, Enterococcus hirae 9790, S. pyogenes MGAS315, S. pneumoniae 262, S. aureus F-182, and Listeria monocytogenes EGD (Fig. 6B and Table 2). In bacteria with l-Ala-l-Ala-cross-bridging peptidoglycans, including E. faecalis strains OG1S, OG1X, OG1RF, and FA2-2 and S. pyogenes, BacL₁ clearly localized in the equatorial ring (38, 43–45). In contrast, the BacL₁ signal was not detected on E. faecium, E. hirae, or S. aureus, which have l-Asp-, d-Asn-, or penta-Gly-cross-bridging peptidoglycan, respectively (37, 46, 47). L. monocytogenes, which has direct bridging between stem peptides, was also not bound with BacL₁ (48, 49). In the case of S. pneumoniae, which has a hetero-cross-bridging structure consisting of l-Ala-l-Ala and l-Ala-l-Ser, the equatorial localization was not observed; however, localization in the division septum was detected (50). Collectively, these observations suggest that BacL₁ specifically binds to the l-Ala-l-Ala-cross-bridged peptidoglycan. On the other hand, the bactericidal phenotype of BacL₁ and BacA was observed only in E. faecalis strains and not in the other bacterial species in soft-agar bacteriocin assays (Table 2). It is notable that S. pyogenes and S. pneumoniae were not susceptible to BacL₁ and BacA despite the targeting of BacL₁ to their cell surface (Table 2).

FIG 6.

BacL₁ localization in various Gram-positive bacterial species. (A) Peptidoglycan structure of E. faecalis, representing an example of the organization of peptide chain-cross-linking by a dipeptide. The dotted-line frame indicates the cross-bridging peptide between stem peptides bound to N-acetylmuramic acids. Arrows indicate the sites of cleavage by the endopeptidase activity of BacL₁. (B) Overnight cultures of Gram-positive bacteria, diluted 5-fold with fresh THB broth, were treated with HiLyte Fluor 555-labeled (red) BacL₁ (5 μg/ml). After incubation for 1 h, the cells were fixed and analyzed by fluorescence microscopy. Phase contrast is pseudocolored (green) in the merged images.

TABLE 2.

Summary of cross-bridge structure and phenotypes against Bac41 in various bacterial species

Species	Strain	Cross-bridging peptide	Presence of phenotypea
			Targeting of BacL1b	Susceptibility to Bac41c
Enterococcus faecalis	OG1S	l-Ala-l-Ala	+	+
Enterococcus faecalis	OG1X	l-Ala-l-Ala	+	+
Enterococcus faecalis	OG1RF	l-Ala-l-Ala	+	+
Enterococcus faecalis	FA2-2	l-Ala-l-Ala	+	±
Enterococcus faecium	BM4105RF	l-Asp	−	−
Enterococcus hirae	9790	d-Asn	−	−
Streptococcus pyogenes	MGAS315	l-Ala-l-Ala	+	−
Streptococcus pneumoniae	262	l-Ala-l-Ala/l-Ser	±	−
Staphylococcus aureus	F-182	Gly5	−	−
Listeria monocytogenes	EGD	NAd	−	−

^a+, clear/positive; ±, obscure/weak; −, negative.

^bTargeting of BacL₁ was determined from the results shown in Fig. 6B.

^cSusceptibility to Bac41 (BacL₁ and BacA mixture) was determined by a soft-agar-based bacteriocin assay.

^dNA, not applicable; L. monocytogenes has direct bridging between stem peptides.

Immunity factor does not alter the BacL₁ equatorial targeting.

The BacL₁- and BacA-producing E. faecalis has a self-resistance factor, BacI, encoded in the vicinity of the bacA gene (Fig. 1A) (26). E. faecalis carrying the bacI gene is completely resistant to the bacteriolytic effect of BacL₁ and BacA (26, 28). Therefore, by fluorescence microscopy, we investigated whether the immunity factor bacI affects the BacL₁ targeting. The equatorial localization of BacL₁ was observed in E. faecalis carrying pHT1100 (a plasmid containing all Bac41 genes, including immunity factor bacI), as well as in E. faecalis carrying pAM401 (a vector control without the bacI gene) (Fig. 7A). Furthermore, the peptidoglycan purified from E. faecalis carrying pHT1100 was still degraded by BacL₁ (Fig. 7B). These results suggest that the specific immunity factor, BacI, has no effect on the BacL₁ activities of binding, targeting, and degrading peptidoglycan.

FIG 7.

Involvement of Bac41 specific immunity factor BacI in the susceptibility of cell wall to BacL₁. (A) An overnight culture of E. faecalis carrying pAM401 (a vector control without the bacI gene) or pHT1100 (a pAM401 derivative containing all Bac41 genes, including immunity factor bacI) diluted 5-fold with fresh THB broth was treated with HiLyte Fluor 555-labeled (red) BacL₁ (5 μg/ml). After incubation for 1 h, the cells were fixed and analyzed by fluorescence microscopy. Phase contrast (Ph) is pseudocolored (green) in merged images. (B) A cell wall fraction prepared from E. faecalis carrying pAM401 or pHT1100 in exponential phase was diluted with PBS. Recombinant BacL₁ (5 μg/ml) or mutanolysin (1 μg/ml) was added to the cell wall suspension, and the mixture incubated at 37°C. The turbidity at 600 nm was quantified at the indicated times during incubation. The values presented are the percentages of the initial turbidity of the respective samples. The PBS-treated sample is presented in each graph as a negative control. The data are presented as the mean results ± SD of three independent experiments.

DISCUSSION

In this study, we report that BacL₁ targets the cell division-associated site, including the equatorial ring, division septum, and nascent synthesized cell wall (Fig. 2), to exert potential bactericidal activity against E. faecalis cells in the dividing state (Fig. 4 and 5). We also demonstrate that BacL₁ specifically recognizes peptidoglycan structures cross-linked by l-Ala-l-Ala but not by other peptide linkers (Fig. 6). Although the entire cell wall in E. faecalis is composed of l-Ala-lAla-cross-bridged peptidoglycan that is likely to be recognized by BacL₁, there must be an additional determinant(s) for the localized targeting of BacL₁ to the cell division-associated sites. The equatorial ring is a characteristic structure observed at the middle of ovococcus cells (32, 51, 52). This ring structure marks the initiation site for the peripheral cell wall-synthesizing machinery to construct the new peptidoglycan during cell elongation. Therefore, BacL₁ might recognize the cell wall-synthesizing machinery complex that is formed at the equatorial ring or division septum during cell division. Alternatively, the relatively extended distribution of BacL₁, from equatorial ring to division septum, raises the possibility that BacL₁ preferentially binds to newly synthesized nascent cell wall. Martínez et al. demonstrated that a bacteriocin of Lactococcus lactis, lactococcin 972, inhibits the septum formation to cause abnormal cell morphology in sensitive target cells. Although they have not shown this, lactococcin 972 itself might be associated with the cell-dividing structure, like BacL₁ (53). Understanding the determinant(s) restricting the targeting site of BacL₁ to cell division-related areas requires further analysis.

As shown by the results in Fig. 3, the SH3 repeat moiety of BacL₁ was required and sufficient for its localized targeting. These repeats are present in the region from amino acid 329 to 590 of BacL₁ (see Fig. S2A in the supplemental material). These individual SH3 repeats are nearly identical to each other (see Fig. S2B). The SH3 domain sequences of BacL₁ also show significant homology to SH3 domains from other bactericidal proteins (see Fig. S2C), such as ALE-1 from S. aureus (54). Crystal structure analysis of the SH3 domain in ALE-1 revealed that the N-terminal conserved motif YXXNKYGTXYXXESA is a recognition groove that specifically binds to penta-Gly-cross-bridging peptides in S. aureus peptidoglycan (42). The YXXNKYGTXYXXESA motif (see Fig. S2C, blue frames) is not present in the SH3 domain of BacL₁. Instead, extra conserved residues (see Fig. S2C, red frames) are present among the SH3 domains targeting bacteria with an l-Ala-l-Ala-cross-bridged cell wall, including E. faecalis, Streptococcus agalactiae, and S. pneumoniae. Furthermore, amino acids 15 and 14 in the N terminus and C terminus, respectively, are highly conserved motifs (see Fig. S2C, magenta highlighting) among the three SH3 domains of BacL₁, suggesting that these conserved motifs in BacL₁ may play a role in the specific recognition of the l-Ala-l-Ala-cross-bridged peptidoglycan structure.

Lysostaphin, with activity specific against S. aureus, is able to distinguish the penta-Gly-cross-bridging structure in the peptidoglycan of S. aureus from the cross-bridging structures of other peptidoglycans (55). The lysostaphin-specific immunity factor Lif, a FemABX-like protein, incorporates serine into the cross-bridging peptides in peptidoglycan of S. aureus and converts it from the penta-Gly-type cross bridge (56). This conversion of the cross-bridging peptide in peptidoglycan results in resistance to lysostaphin. Zoocin A is a bacteriolytic endopeptidase against the cell wall of sensitive bacteria produced by Streptococcus equi subsp. zooepidemicus strain 4881 (57). The cross bridge in peptidoglycan of S. equi is an l-Ala-l-Ala peptide and is susceptible to the peptidoglycan hydrolase activity of zoocin A (57). Zif, an immunity factor of zoocin A, belongs to the FemABX-like protein family (58). It additively increases l-Ala residues in the cross bridges of peptidoglycans and converts l-Ala-l-Ala into l-Ala-l-Ala-l-Ala, resulting in resistance to zoocin A activity. Meanwhile, BacI, which is the cognate immunity factor against Bac41, did not affect the BacL₁ targeting (Fig. 7A). In addition, the cell wall fraction prepared from E. faecalis that is resistant to Bac41 due to the presence of bacI was still susceptible to the peptidoglycan-degrading activity of BacL₁ (Fig. 7B), suggesting that BacI is not involved in the activity of BacL₁. This result suggests the possibility that BacI confers immunity by acting on the function of BacA rather than that of BacL₁ or that another factor(s) of target cells, such as molecules or receptors that are only present in the growing cells, is involved in the BacI-mediated resistance.

The bactericidal activity of Bac41 (BacL₁ and BacA) is strictly specific against E. faecalis, and Bac41 does not show any activity against the other bacterial species tested (Table 2). The specificity could be partially explained by the diversity of cross-bridging peptides of peptidoglycan among bacterial species. As demonstrated by the results in Fig. 6B, BacL₁ appears to discriminate target bacterial species from nontarget species by specific recognition of l-Ala-l-Ala-cross-bridged peptidoglycan. Indeed, BacL₁ is able to target bacteria with l-Ala-l-Ala-cross-bridged peptidoglycan, such as S. pyogenes and S. pneumoniae, regardless of the bacterial genus. In contrast, E. faecium and E. hirae, with peptidoglycans cross bridged by l-Asp and d-Asn, respectively, were not recognized by BacL₁ although they are phylogenetically classified in the same genus as E. faecalis. These observations demonstrated that the activity of BacL₁ is specific against bacteria with l-Ala-l-Ala-cross-bridged peptidoglycans. However, the bacteriolytic phenotype in the copresence of BacL₁ and BacA appears to be more complex (Table 2). The bactericidal effect of BacL₁ and BacA (Bac41) was observed only against E. faecalis even though other bacteria are of the l-Ala-l-Ala-cross-bridge type. Interestingly, S. pyogenes and S. pneumoniae were not susceptible to BacL₁ and BacA although they were targeted with BacL₁. One possibility is that BacA is not able to access S. pyogenes and S. pneumoniae. Furthermore, the susceptibility of E. faecalis FA2-2 to BacL₁ and BacA was lower than that of E. faecalis OG1-derived strains, such as OG1S, OG1X, and OG1RF. Thurlow et al. reported that enterococcal capsular polysaccharide is present in FA2-2 but not in OG1 strains (59). Thus, probably the capsule on the cell surface of strain FA2-2 cells limits the access of BacA, resulting in the decreased susceptibility to Bac41-induced lysis. To reveal the detailed molecular mechanism of the Bac41 module, further functional analysis of BacA is needed.

The Bac41-mediated fratricide module excludes E. faecalis strains without the Bac41-encoding plasmid. Therefore, this module is inferred to play a role in the effective expansion of the Bac41-carrying plasmid. Our conclusion that cell growth is required for cell lysis by BacL₁ and BacA (Fig. 4 and 5) is consistent with the hypothesis because selection is involved in possible plasmid loss during distribution to daughter cells. Hence, it is reasonable that the Bac41 system works only when bacteria are allowed to grow, replicate DNA, and distribute plasmid to daughter cells. Our results in this study suggest a novel player involved in the plasmid maintenance system.