The Cationic Antimicrobial Peptide Activity of Lysozyme Reduces Viable Enterococcus faecalis Cells in Biofilms

ABSTRACT

Enterococcus faecalis, a leading cause of health care-associated infections, forms biofilms and is resistant to many antimicrobial agents. Planktonic-phase E. faecalis is resistant to high concentrations of the enzyme lysozyme, which catalyzes the hydrolysis of N-acetylmuramic acid and N-acetylglucosamine linkages in peptidoglycan and is also a cationic antimicrobial peptide (CAMP). E. faecalis lysozyme resistance in planktonic cells is stimulated upon activation of the extracytoplasmic function sigma factor SigV via cleavage of the anti-sigma factor RsiV by the transmembrane protease Eep. Planktonically grown E. faecalis lacking eep is more sensitive than wild-type strains to growth inhibition by lysozyme. This study was initiated to determine whether E. faecalis OG1RFΔeep biofilms would be protected from lysozyme. Serendipitously, we discovered that exposure of both E. faecalis OG1RF and OG1RFΔeep biofilms to chicken egg white lysozyme resulted in decreases in biofilm cell viability of 3.7 and 3.8 log₁₀ CFU/mL, respectively. Treatment of biofilms of both strains with recombinant purified human lysozyme was associated with reductions in cell viability of >99.9% for both strains. Lysozyme-treated OG1RF and OG1RFΔeep biofilms contained a higher percentage of dead cells by Live/Dead staining and were associated with more extracellular DNA. Heat-inactivated human lysozyme, which was devoid of muramidase activity, as well as the lysozyme-derived CAMP LP9 and the CAMP polymyxin B, decreased biofilm cell viability. These results are consistent with a model in which the CAMP activity, rather than the muramidase activity, of lysozyme causes lysis of E. faecalis biofilm cells despite them having an intact lysozyme resistance-inducing signaling pathway. Finally, lysozyme was also effective in reducing viable biofilm cells of several other E. faecalis strains, including the vancomycin-resistant strain V583 and multidrug-resistant strain MMH594. This study demonstrates the potential for lysozyme to be developed as a novel antibiofilm therapeutic.

INTRODUCTION

Enterococcus faecalis, a natural microbial inhabitant of the human gastrointestinal tract, has emerged as a leading cause of health care-associated infections (1, 2). E. faecalis infections include wound infections, endocarditis, catheter-associated urinary tract infections (CAUTI), and bloodstream infections (1). Treatment options for these infections have become increasingly limited due to both the intrinsic and acquired antibiotic resistance traits of enterococcal strains (3, 4). Further, E. faecalis is able to persist in a range of environmental conditions, such as high temperatures and low pH (5, 6). The ability of E. faecalis to withstand high levels of stress is also observed in the presence of host-derived antimicrobial molecules, particularly in response to the cell envelope-targeting enzyme lysozyme (7, 8). Lysozyme is a component of the innate immune system that cleaves the β-1,4-linkages between the N-acetylmuramic acid and N-acetylglucosamine residues of peptidoglycan, leading to cell lysis. Lysozyme can also act as a cationic antimicrobial peptide (CAMP) that destabilizes the bacterial cell membrane (9).

E. faecalis is inherently resistant to the bactericidal activity of lysozyme (10). Lysozyme resistance in E. faecalis involves activation of the extracytoplasmic sigma factor SigV via the proteolytic degradation of the corresponding anti-sigma factor RsiV through a process called regulated intramembrane proteolysis (5, 7, 11, 12). In the absence of lysozyme, SigV is rendered inactive by the membrane-anchored RsiV. After lysozyme exposure, RsiV is sequentially cleaved by a site-1 protease at the extracytoplasmic domain and by a site-2 protease at the transmembrane domain (11–13). The site-2 protease responsible for the liberation of SigV in E. faecalis is a transmembrane metalloprotease called Eep (11). SigV is resultantly released into the cytosol, where it interacts with RNA polymerase to induce genes that confer resistance to lysozyme. Genes that promote modification of the cell envelope (oatA, dltA, and pgdA) have been shown to influence lysozyme resistance in E. faecalis, although the extent to which this occurs varies on the strain background (7, 8, 11, 14). In E. faecalis strains V583 and FA2-2, deletion of eep increased susceptibility to lysozyme more than 12.5-fold compared to the wild type (WT), stressing the critical role of Eep in lysozyme resistance (11).

The Eep-mediated lysozyme-resistance pathway was elucidated in E. faecalis using planktonic broth-based assays (11). Despite the critical role of biofilm formation in enterococcal infection (15), little is known about the effect of lysozyme on E. faecalis cells in the biofilm state. Enterococci in biofilms have been shown to be more resistant to antimicrobial agents than their planktonic counterparts (16–18). In addition, Eep—which promotes lysozyme resistance in planktonic cells—is also important for E. faecalis biofilm formation (19, 20). We reported previously that a mutant strain lacking Eep (OG1RFΔeep) exhibited altered biofilm cell and matrix distribution compared to the wild type in the early stages of biofilm growth (20). Furthermore, the virulence of OG1RFΔeep was attenuated in experimental animal models for endocarditis and CAUTI, demonstrating that Eep is necessary for pathogenesis in biofilm-associated enterococcal infections (19, 20).

Since enterococcal biofilms offer a protective advantage against high concentrations of antimicrobial agents (16), we initially set out in this work to test the hypothesis that E. faecalis OG1RFΔeep biofilms would be protected from lysozyme. Instead, we found unexpectedly that biofilms of the parent strain, OG1RF, were highly susceptible to killing by lysozyme. Moreover, the viability of cells in OG1RF WT and OG1RFΔeep biofilms was equally reduced after exposure to lysozyme, indicating that Eep does not protect biofilm cells against the bactericidal activity of lysozyme. We also demonstrate that treatment of E. faecalis biofilms with CAMPs is an effective method of inducing cell lysis. Importantly, this is the first study to identify lysozyme as an agent with therapeutic potential against multidrug-resistant enterococcal biofilms.

RESULTS

Eep is necessary for growth of E. faecalis OG1RF in the presence of lysozyme under nonbiofilm conditions.

Eep protease confers resistance to lysozyme in the E. faecalis strains V583 and FA2-2 (11). To determine whether Eep similarly contributes to lysozyme resistance in E. faecalis OG1RF, serial dilutions of planktonically grown broth cultures of the WT and isogenic Δeep strains were plated on increasing concentrations of chicken egg white (CEW) lysozyme. OG1RF was fully able to grow in the presence of 5 mg/mL lysozyme (Fig. 1A). In contrast, growth of OG1RFΔeep was suppressed between 4 to 5 orders of magnitude at ≥1.5 mg/mL lysozyme. Planktonic MICs of lysozyme after 20 h of growth were 12.5 and 0.8 mg/mL for OG1RF and OG1RFΔeep, respectively (see Fig. S1 in the supplemental material). These findings verify that the OG1RFΔeep mutant is more susceptible than the parental strain to growth inhibition by lysozyme under nonbiofilm conditions.

FIG 1.

Deletion of the eep ORF in E. faecalis OG1RF results in the inhibition of logarithmic phase growth under planktonic culture conditions. (A) Overnight planktonic cultures of OG1RF and OG1RFΔeep were serially diluted and plated on BHI agar containing the indicated concentrations of lysozyme. Each data point is an independent biological replicate. Horizontal bars indicate the geometric means. (B and C) CEW lysozyme (5 mg/mL in water) or water only was added to early logarithmic-phase (B) or stationary-phase (C) planktonic cells. Cell viability was measured at 0, 1, 3, and 6 h postexposure. Values at each time point represent the means ± the standard deviations of three independent replicates.

Interestingly, punctate forms were observed repeatedly with the OG1RFΔeep strain in MIC plate wells containing 0.8 to 6.25 mg/mL of lysozyme (see Fig. S1). Lysozyme concentrations of ≥12.5 mg/mL inhibited the formation of the punctate forms. Subsequent culture of the broth in the indicated wells (see Fig. S1) suggested that the punctate forms were spontaneous mutants which suppress the OG1RFΔeep lysozyme sensitivity phenotype (data not shown). Our group has isolated similar suppressor mutants in other conditions and, in separate work, has characterized such mutants in detail (C. N. Rouchon and K. L. Frank, unpublished data); therefore, the suppressor mutants isolated in Fig. S1 were not further studied in this work.

The E. faecalis lysozyme resistance phenotype is specific to growth phase.

To measure the effect of bacterial growth phase on Eep-mediated lysozyme resistance, logarithmic- and stationary-phase planktonic E. faecalis cells were cultured in the presence or absence of 5 mg/mL CEW lysozyme for 6 h. Logarithmic-phase OG1RF cell numbers increased by ~1 log₁₀ CFU/mL at 6 h postincubation (Fig. 1B), confirming that the WT strain was resistant to killing by lysozyme during this growth stage. The number of viable logarithmic-phase OG1RFΔeep cells decreased from 1.6 × 10⁷ CFU/mL to 5 × 10³ CFU/mL after 6 h of treatment (Fig. 1B). These findings indicate that eep gene is essential for lysozyme resistance in actively growing cells. In contrast, exposure of stationary-phase cultures to lysozyme over the same time period resulted in a reduction in viability from ~1.1 × 10⁷ CFU/mL to ~3.5 × 10⁵ to 7 × 10⁵ CFU/mL for both OG1RF and OG1RFΔeep (Fig. 1C), demonstrating that OG1RF and OG1RFΔeep are similarly sensitive to lysozyme at later stages of growth. Together, the data confirm that E. faecalis lysozyme resistance is dependent upon the growth phase of the cells.

Lysozyme treatment of E. faecalis biofilms equally reduces the number of viable OG1RF and OG1RFΔeep cells.

Twenty-hour-old E. faecalis biofilms were exposed to 5 mg/mL CEW lysozyme for 3 h and then assayed for biofilm biomass and cell viability. Unexpectedly, we observed an increase in stainable biomass for both OG1RF and OG1RFΔeep lysozyme-treated biofilms relative to biofilms exposed to buffer only (Fig. 2A). Similar results were obtained when biofilms were stained with crystal violet (see Fig. S2A), indicating that the observed increase in biomass was not a stain-specific phenomenon. The lysozyme-associated increase in stained biomass shown in Fig. 2A correlated with reductions in bacterial cell survival of 3.7 log₁₀ CFU/mL for OG1RF and 3.8 log₁₀ CFU/mL for OG1RFΔeep (Fig. 2B). Thus, in this assay, we observed >99.9% killing for both strains after treatment with CEW lysozyme. No such effects occurred when OG1RF and OG1RFΔeep biofilms were exposed to the peptidoglycan-targeting antibiotic ampicillin (Fig. 2C and D). Ampicillin is bactericidal against planktonic OG1RF (MIC = 0.5 μg/mL; MBC = 1 μg/mL) but is ineffective against OG1RF biofilms (biofilm bactericidal concentration > 128 μg/mL) and also, unlike lysozyme (Fig. 1C), static-phase cells (17; data not shown). In addition, there was no alteration to the biomass or the viability of OG1RF and OG1RFΔeep biofilms exposed to bovine serum albumin, a nonenzymatic protein used as a control (see Fig. S2B and C).

FIG 2.

Lysozyme reduces the number of viable E. faecalis OG1RF and OG1RFΔeep cells in biofilms. (A to F) Biofilms grown for 20 h in 96-well microtiter plates were exposed to 5 mg/mL CEW lysozyme (A and B), 128 μg/mL ampicillin (C and D), or 1.25 mg/mL purified recombinant human lysozyme (E and F). Control biofilms were exposed to 10 mM Tris-HCl buffer (pH 8) or water only. Biofilm biomass and bacterial cell survival were assessed 3 h postexposure. The data are presented as means ± the standard deviations of three biological replicates. The dotted line in panel B indicates the limit of detection for the supernatant samples. b.d.l., below detection limit. Statistical significance was calculated using two-way ANOVA with Sidak’s multiple-comparison test (****, P < 0.0001).

We next determined if lower concentrations of CEW lysozyme or longer treatment times could further reduce cell viability in E. faecalis biofilms (Table 1 and Fig. S3, respectively). Exposure of biofilms to 0.16 mg/mL lysozyme decreased bacterial cell survival by 2.6 (>99%) and 3.3 (>99.9%) log₁₀ CFU/mL for OG1RF and OG1RFΔeep, respectively, relative to controls (Table 1). Treatment of biofilms with higher concentrations of lysozyme resulted in similar levels of killing, with reductions of ~2.8 to 3.7 log₁₀ CFU/mL (Table 1). Quantification of the biofilm cell viability following exposure to the highest concentration of lysozyme (5 mg/mL) for 3, 6, or 24 h revealed that the majority of killing (>99.9%) occurred by 3 h posttreatment for both OG1RF and OG1RFΔeep (see Fig. S3).

TABLE 1.

Effect of lysozyme concentration on biofilm biomass and cell viability

Strain	Treatment	Mean biofilm biomass and log10 CFU values (±SD) at various lysozyme concentrations
		0.16 mg/mL		1.25 mg/mL		5 mg/mL
		Biofilm biomass (OD450)	Log10 CFU/mL	Biofilm biomass (OD450)	Log10 CFU/mL	Biofilm biomass (OD450)	Log10 CFU/mL
OG1RF	Buffer	0.09 ± 0.03	7.89 ± 0.14	0.09 ± 0.03	7.85 ± 0.13	0.11 ± 0.02	7.97 ± 0.10
	Lysozyme	0.19 ± 0.08	5.29 ± 0.55	0.28 ± 0.00	4.11 ± 0.59	0.20 ± 0.05	5.21 ± 0.96
OG1RFΔeep	Buffer	0.09 ± 0.02	8.14 ± 0.44	0.09 ± 0.03	7.66 ± 0.14	0.12 ± 0.01	8.00 ± 0.07
	Lysozyme	0.18 ± 0.06	4.84 ± 0.52	0.28 ± 0.01	4.14 ± 0.53	0.22 ± 0.09	4.98 ± 0.82

Treatment of E. faecalis biofilms with recombinant human lysozyme corresponds with bactericidal levels of reduced cell viability.

E. faecalis is a member of the microbiota in the human gastrointestinal tract, which contains lysozyme-producing Paneth cells (21); therefore, we determined whether purified, recombinant human lysozyme exhibited antibiofilm activity. Similar to the results with CEW lysozyme, treatment of OG1RF and OG1RFΔeep biofilms with 1.25 mg/mL of human lysozyme resulted in an ~3.5-fold increase in stained biomass compared to buffer-only-treated biofilms (Fig. 2E). Human lysozyme also reduced the number of viable cells recovered from OG1RF and OG1RFΔeep biofilms by 3.7 and 3.8 log₁₀ CFU/mL, respectively (Fig. 2F), which are decreases of 99.98%. Therefore, in contrast to the lysozyme-resistant phenotype demonstrated by planktonic phase OG1RF (Fig. 1), lysozyme from both chicken egg whites and humans reduces the number of viable cells in biofilms formed by this E. faecalis strain.

Lysozyme is bactericidal against E. faecalis biofilm cells.

The data in Fig. 2 suggest that exposure of E. faecalis OG1RF and OG1RFΔeep biofilms to lysozyme may cause either dispersal or lysis of biofilm cells. To determine whether lysozyme treatment caused dispersal of live bacteria from established biofilms, the number of viable cells present in the supernatant following lysozyme or buffer-only treatment in Fig. 2A and B was quantified. Supernatant collected from the control wells contained averages of 6.8 and 6.7 log₁₀ CFU/mL for OG1RF and OG1RFΔeep, respectively, indicating that viable cells were dispersed from the biofilms (Fig. 2B). However, supernatant collected from lysozyme-treated biofilms contained no detectable bacterial cells. These data indicate that the reduction in viable biofilm cells following lysozyme treatment is not simply due to biofilm cell dispersal.

Cell viability of lysozyme-treated E. faecalis biofilms was assessed with BacLight Live/Dead fluorescence staining (Fig. 3). Lysozyme-treated biofilms contained a higher percentage of propidium iodide-stained cells (red) compared to buffer-only-treated controls, indicating that these cells have compromised membranes and are likely dead. These images independently corroborate the finding that exposure of biofilms to lysozyme leads to the reduction in the number of viable E. faecalis cells recovered (Fig. 2B and F). Furthermore, the combined findings from Fig. 2 and 3 strongly suggest that the reduction in viable E. faecalis biofilm cells observed following lysozyme exposure is due to cell lysis.

FIG 3.

Exposure of E. faecalis OG1RF and OG1RFΔeep biofilms to lysozyme leads to increased cell death. Biofilms cultivated on Aclar coupons were exposed to 5 mg/mL CEW lysozyme or buffer only for 3 h. Treated biofilms were labeled with Live/Dead stain solution and visualized by confocal laser scanning microscopy to identify live (green; Syto9) and dead (red; propidium iodide) cells. Representative images are maximum intensity projections of biofilm z-stacks visualized with a 63× oil objective. Experiments were performed in duplicate from independent biological replicates. Scale bars, 10 μm.

Lysozyme-treated biofilms are associated with increased amounts of extracellular DNA.

Lysozyme-mediated lysis of the bacterial biofilm cells may result in the release of cellular DNA into the extracellular milieu. To test this, PicoGreen dsDNA quantitation reagent was added to biofilms treated with 5 mg/mL CEW lysozyme or buffer only. Lysozyme treatment was associated with a ≥3.5-fold increase in fluorescence (Fig. 4A) compared to buffer-only controls. The increase in fluorescence correlated with decreases in bacterial cell viability of 3.9 and 4.1 log₁₀ CFU/mL for OG1RF and OG1RFΔeep, respectively (Fig. 4B). The application of 10 U/mL of micrococcal nuclease to lysozyme-treated biofilms reduced fluorescence by >3-fold compared to mock-treated controls (Fig. 4C), verifying that biofilms exposed to lysozyme contained more extracellular (eDNA) than buffer-only biofilms. In addition, the optical density at 450 nm (OD₄₅₀) of lysozyme-treated biofilms decreased by ~4.5-5-fold after the addition of 10 U/mL of micrococcal nuclease (Fig. 4D). Although we cannot rule out the possibility that the increase in stainable biomass observed after lysozyme treatment (Fig. 2A and E) was due, in part, to lysozyme binding to the biofilm cells, the data in Fig. 4D strongly suggest that the increase in stained biofilm biomass was most likely due to the presence of DNA released from lysed biofilm cells.

FIG 4.

Lysozyme-treated E. faecalis biofilms have larger amounts of eDNA. (A) PicoGreen, a DNA-binding fluorescent dye, was added to washed biofilms after a 3-h exposure to lysozyme or buffer in order to detect and quantify eDNA. (B) Viability of cells manually released from biofilms in panel A and enumerated on BHI agar. (C and D) The PicoGreen fluorescence (C) and biofilm biomass (D) were assessed after a 1-h application of 10 U/mL of micrococcal nuclease to lysozyme-treated biofilms. The results are reported as means ± the standard deviations of three biological replicates. Statistical significance was calculated using two-way ANOVA with Sidak’s multiple-comparison test (****, P < 0.0001).

Genes involved in E. faecalis planktonic lysozyme resistance are not involved in the lysozyme-mediated killing of biofilm cells.

Lysozyme resistance in planktonically grown E. faecalis cells is associated with the expression of cell envelope-modifying genes such as pgdA, oatA, and dltA (7, 8, 11, 14). Of these, only pgdA is activated by SigV in the signaling pathway that involves Eep cleavage of SigV’s anti-sigma factor, RsiV, after sensing of lysozyme at the cell envelope (7, 14, 22). Since E. faecalis biofilms were sensitive to lysozyme (Fig. 2), we compared the expression of lysozyme resistance genes between biofilm and planktonic cells. Expression of dltA, oatA, pgdA, rsiV, and sigV was measured in OG1RF and OG1RFΔeep cells from corresponding planktonic and biofilm cultures after 6 h of growth; insufficient expression (C_T values ≥ 30) of two potential reference genes (relA and gyrB) in cells from 18-h-old biofilms precluded analysis at this later time point. There were no fold change expression levels of >2 when comparing planktonic to biofilm cells for either OG1RF or OG1RFΔeep (see Fig. S4A).

We also tested the activity of lysozyme against biofilms of OG1RFΔsigV and OG1RFΔeepΔsigV in-frame deletion mutant strains, which are susceptible to the same concentration of lysozyme as OG1RFΔeep (see Fig. S1B and C). Lysozyme treatment of OG1RFΔsigV and OG1RFΔsigVΔeep biofilms caused increases in biomass (see Fig. S4B), similar to what was observed for OG1RF and OG1RFΔeep (Fig. 2A; see also Fig. S4A). Cell viability reductions of similar magnitude for all strains were also observed following lysozyme treatment (see Fig. S4C). Together, these results suggest that lysozyme-mediated killing of E. faecalis biofilm cells occurs independently of the enterococcal planktonic lysozyme resistance signaling pathway.

Mutanolysin, an alternate muramidase, is also active against E. faecalis biofilms.

To better understand how muramidase activity may contribute to the killing of E. faecalis biofilm cells, we next investigated the antibiofilm activity of an alternate muramidase enzyme, mutanolysin (23). Unlike lysozyme (Fig. 1), mutanolysin had no effect on the viability of planktonically grown E. faecalis cells in either the logarithmic or the stationary phases of growth (Fig. 5A and B). Exposure of OG1RF and OG1RFΔeep biofilms to 500 U/mL of mutanolysin caused an increase in stainable biomass compared to buffer-only-treated samples for only OG1RF (Fig. 5C). Interestingly, mutanolysin treatment reduced the number of viable cells recovered from biofilms by 2.8 log₁₀ CFU/mL for OG1RF and 2.7 log₁₀ CFU/mL for OG1RFΔeep, relative to buffer controls (Fig. 5D), but ca. 2.5 to 3 log₁₀ CFU/mL viable cells were detected in the supernatants of mutanolysin-treated biofilms (Fig. 5D). Since no viable cells were detected in the supernatants of biofilms exposed to lysozyme (Fig. 2B), these data indicate that the muramidase mutanolysin works differently from lysozyme to decrease the number of viable cells in E. faecalis cells in biofilms.

FIG 5.

Mutanolysin, an alternate muramidase, reduces the viability of E. faecalis biofilm cells but does not affect planktonic or dispersed biofilm cells. (A and B) Normalized cultures (OD₆₀₀ ~ 0.020) of early logarithmic-phase (A) or stationary-phase (B) planktonic cells were exposed to mutanolysin (500 U/mL in water) or water only at 37°C. Cell viability was enumerated at 0, 1, 3, and 6 h posttreatment. The data represent the means ± the standard deviations of two biological replicates. (C and D) Twenty-hour-old biofilms were cultivated in microtiter plates and then treated with 500 U/mL mutanolysin. Control samples were exposed to water only. The data show the means ± the standard deviations of three biological replicates. The dotted line in panel D indicates the limit of detection for the supernatant samples. Statistical significance was calculated using two-way ANOVA with Sidak’s multiple-comparison test (*, P < 0.05; ****, P < 0.0001 compared to control).

CAMP activity alone is sufficient to mediate the antimicrobial effects of lysozyme on E. faecalis biofilms.

Lysozyme possesses both enzymatic muramidase activity that targets cell wall peptidoglycan and CAMP properties that destabilize the bacterial membrane (24, 25). To determine whether the muramidase activity of lysozyme was required to disrupt E. faecalis biofilms, human recombinant lysozyme was boiled for 1 h at 100°C and then applied to 20 h-old biofilms. Heat-treated lysozyme that was subsequently incubated at 37°C for up to 6 h did not lyse Micrococcus lysodeikticus cells, verifying that the denatured enzyme was catalytically inactive and had not refolded (see Fig. S5). Both heat-inactivated (HI) lysozyme and recombinant human lysozyme treatment caused an increase in biofilm biomass for OG1RF and OG1RFΔeep, in comparison buffer-treated biofilms (Fig. 6A). Relative to control samples, application of recombinant human lysozyme decreased biofilm cell numbers by 4.8 and 4.3 log₁₀ CFU/mL for OG1RF and OG1RFΔeep, respectively (Fig. 6B). Treatment with HI lysozyme also reduced biofilm cell viability by >4 log₁₀ CFU/mL for both strains (Fig. 6B). These findings demonstrate that lysozyme is capable of damaging E. faecalis biofilms in the absence of muramidase activity.

FIG 6.

CAMP activity is sufficient to reduce the survival of E. faecalis cells in biofilms. Twenty-hour-old biofilms were cultivated in microtiter plates and then treated with 1.25 mg/mL recombinant human lysozyme or heat-inactivated human lysozyme (A and B) or 1.25 mg/mL human lysozyme, 200 μg/mL LP9 peptide, or 200 μg/mL polymyxin B (C and D). Control samples were exposed to 10 mM Tris-HCl (pH 8) or water only. The means ± standard deviations of ≥2 biological replicates are presented. Statistical significance was calculated using two-way ANOVA with Dunnett’s multiple-comparison test (****, P < 0.0001 compared to control).

HI lysozyme has previously been shown to retain CAMP activity (24, 26). Therefore, we determined if the peptide LP9, a CAMP derived from human lysozyme, had an effect on biofilm cell survival. Exposure of biofilms to 200 μg/mL of LP9 did not alter biomass (Fig. 6C). However, although the effect was not as potent as lysozyme or HI lysozyme, the LP9 peptide reduced biofilm cell viability by nearly 2 log₁₀ CFU/mL (>90%) for OG1RF and OG1RFΔeep compared to buffer alone (Fig. 6D). Further, treatment of biofilms with 200 μg/mL of polymyxin B, a CAMP that is not related to lysozyme, decreased biofilm cell numbers by 3.4 log₁₀ CFU/mL relative to buffer only controls (Fig. 6D). Biofilm biomass was not affected by polymyxin B treatment (Fig. 6C). Finally, neither polymyxin B nor HI lysozyme were active against stationary-phase E. faecalis OG1RF (see Fig. S6). Taken together, these data indicate that CAMP activity is sufficient to kill E. faecalis biofilm cells. However, this conclusion does not extend to stationary-phase E. faecalis cells.

Lysozyme kills biofilms of E. faecalis laboratory strains and clinical isolates.

Finally, we evaluated the biofilm cell-killing effect of lysozyme against biofilms formed by several laboratory strains and clinical isolates of E. faecalis (Table 2), including the vancomycin-resistant isolate V583 (27) and the multidrug-resistant E. faecalis strain MMH594 (28). The strains tested in this study also possessed various biofilm-forming capabilities. CEW lysozyme treatment resulted in an increase in stained biomass of ca. 2- to 2.5-fold and a concomitant decrease in cell viability of ca. 2.8 to 3.5 log₁₀ CFU/mL (>99 to 99.9%) relative to controls (Fig. 7). In summary, the bactericidal activity of lysozyme against biofilms is applicable across E. faecalis isolates, including strains that exhibit resistance to conventional antibiotics.

TABLE 2.

E. faecalis strains used in this study

Strain	Description	Source or reference
OG1RF	Oral isolate, plasmid-free	46
OG1RFΔeep	Markerless in-frame deletion of locus OG1RF_11819 (previously called JRC106)	42
OG1RFΔsigV	Markerless in-frame deletion of locus OG1RF_12448 in strain OG1RF	This study
OG1RFΔeepΔsigV	Markerless in-frame deletion of locus OG1RF_12448 in strain OG1RFΔeep	This study
DS16	Clinical isolate	47
MMH594	Clinical isolate, multidrug resistant	28
V583	Blood isolate, vancomycin resistant	27
VA1128	Clinical isolate	48

FIG 7.

The viability of E. faecalis laboratory strains and clinical isolates is reduced in biofilms after exposure to lysozyme. Biofilms grown in microtiter plates for 20 h were treated with 5 mg/mL CEW lysozyme or buffer only for 3 h. Biofilm biomass (A) and bacterial cell survival (B) were measured for each strain. The data are presented as means ± the standard deviations of three biological replicates. Statistical significance was calculated using two-way ANOVA with Sidak’s multiple-comparison test (*, P < 0.05; ****, P < 0.0001).

DISCUSSION

Lysozyme is a naturally occurring antimicrobial enzyme found in the mucosal secretions of animals and humans, where it occurs in tears at concentrations of ~1.5 mg/mL (29, 30). CEW lysozyme is Generally Regarded As Safe (GRAS) by the U.S. Food and Drug Administration and is approved in the European Union for use as a food additive to control microbial growth. We describe here the novel discovery that lysozyme exerts bactericidal activity against preformed biofilms of multiple E. faecalis strains (Fig. 2, 3, and 7) in a matter of hours (see Fig. S3). Although our efforts to purify a catalytic-inactive mutant of recombinant human lysozyme were unsuccessful (data not shown), our data using heat-inactivated lysozyme suggest that the CAMP activity of lysozyme alone is sufficient for the observed antibiofilm effect (Fig. 6).

The susceptibility of biofilms to lysozyme described here contrasts with the high levels of resistance previously reported for planktonically grown E. faecalis cells following lysozyme exposure (5, 7, 8, 11). Modification of the enterococcal cell wall plays an important role in preventing the bactericidal activity of lysozyme in planktonic cells. O-acetylation and deacetylation of peptidoglycan by OatA and PgdA, respectively, protect the cell wall from the muramidase activity of lysozyme, and d-alanylation of peptidoglycan by Dlt inhibits the cationic properties of the antimicrobial enzyme (7, 8, 11). In B. subtilis, these cell wall-modifying genes are activated through a signaling cascade that is initiated by the binding of lysozyme to RsiV (13, 31, 32). Once lysozyme is bound, RsiV is cleaved by site-1 and site-2 proteases, which ultimately leads to the upregulation of genes such as pgdA. Since E. faecalis recombinant RsiV has also been shown to bind lysozyme, it has been hypothesized that E. faecalis senses lysozyme in a similar manner, leading to degradation of RsiV in part by the site-2 protease Eep with subsequent activation of SigV and cell envelope-modifying genes (13, 14). Our data demonstrate that OG1RFΔeep, OG1RFΔsigV, and OG1RFΔeepΔsigV mutants (Fig. 1; see also Fig. S4B and C) are susceptible to much lower concentrations of lysozyme than WT under planktonic growth conditions (Fig. 1), which is concordant with other E. faecalis strains (11). However, we observed that the viability of OG1RF, OG1RFΔeep, OG1RFΔsigV, and OG1RFΔeepΔsigV biofilm cells was reduced after treatment with lysozyme (Fig. 2; see also Fig. S4C). Furthermore, the genes associated with lysozyme resistance in planktonic cells are not differentially expressed in early biofilm cells in the absence of lysozyme (see Fig. S4). Therefore, an additional major finding from this work is that Eep protease does not confer resistance to lysozyme when E. faecalis cells are in biofilms and suggests that the response of E. faecalis to lysozyme differs between the planktonic and biofilm states.

The data in Fig. 5 and 6 indicate that it is the CAMP activity rather than the muramidase activity of lysozyme which mediates the biofilm-specific killing of E. faecalis. In addition, the number of viable cells recovered from E. faecalis OG1RF biofilms also decreased following incubation with the alternate CAMP polymyxin B (Fig. 6D). Growth phase-dependent susceptibility to lysozyme and other cationic antimicrobial agents in planktonic phase cells has been described for multiple bacterial species. Under specific growth conditions, Lactobacillus fermenti is more resistant to lysozyme in stationary phase than during exponential growth (33). In contrast, methicillin-resistant Staphylococcus aureus (MRSA) clinical strains are more susceptible to CAMPs in the stationary phase of growth compared to logarithmic cultures (34). The reduced viability of stationary-phase MRSA cells following treatment with CAMPs was attributed to changes in the cell surface charge during that phase of growth (34). Similar to these MRSA results, stationary-phase cells of the E. faecalis OG1RF strain were killed by lysozyme more readily than logarithmic-phase cells were (Fig. 1B and C). Mutanolysin, the other murolytic enzyme that we tested, does not have known CAMP activity and did not have an effect on logarithmic or stationary-phase cells (Fig. 5A and B). Stationary-phase E. faecalis OG1RF cells were also not affected by heat-inactivated recombinant human lysozyme or polymyxin B (see Fig. S6). Therefore, CAMP-mediated killing of E. faecalis biofilm cells cannot be simply attributed to the presence of stationary-phase cells within the biofilm. It is known that the E. faecalis OG1RF biofilm matrix contains eDNA, polysaccharides, and proteins (35, 36). One possible explanation for the activity of CAMPs against E. faecalis biofilm cells is that negatively charged macromolecules within the biofilm matrix may attract CAMPs through electrostatic interactions.

Vancomycin-resistant enterococci (VRE) are responsible for approximately 35% of enterococcal infections in the United States and are on the WHO high-priority list of pathogens for which new treatment strategies are needed (37, 38). Options for treatment of biofilm-associated enterococcal infections, particularly ones caused by antibiotic-resistant isolates like VRE, have become more limited (15). Our work demonstrates that lysozyme was effective against the biofilms of a diverse group of E. faecalis strains with various antibiotic resistance and virulence profiles (Fig. 7). Notably, biofilms of the vancomycin-resistant E. faecalis isolate V583 and the multidrug-resistant strain MMH594 were highly susceptible to killing by lysozyme. Therefore, the reduction in biofilm cell viability following exposure to lysozyme is not strain specific.

The bactericidal effect of lysozyme on biofilm cells is not limited to E. faecalis. Thellin et al. reported that treatment of 24-h-old biofilms of Gardnerella vaginalis, the causative agent of bacterial vaginosis, with recombinant human lysozyme significantly decreased the biofilm biomass (39). Lysozyme was also effective against biofilms of G. vaginalis cocultured with Lactobacillus spp., Bacteroides vulgatus, Peptostreptococcus tetradius, or Candida albicans (39). Importantly, lysozyme concentrations up to 100,000 U/mL had no toxic effect on a vaginal epithelial cell line (39). The viability of Staphylococcus epidermidis biofilm cells was also reduced following cotreatment with 16 mg/mL lysozyme and 1 mg/mL lactoferrin (40). Sheffield et al. demonstrated that both Escherichia coli and Klebsiella pneumoniae subsp. pneumoniae biofilms were susceptible to lysozyme (41). Specifically, exposure of 24-h-old E. coli and K. pneumoniae subsp. pneumoniae biofilms with 5 to 50 μg/mL of lysozyme reduced the biomass by ≥73 and 100%, respectively (41). The efficacy of lysozyme against biofilms of a variety of bacterial species demonstrates the potential for the antimicrobial enzyme to be used as a broad-spectrum antibiofilm therapeutic.

In conclusion, we have shown that the innate immune molecule lysozyme is a potent killer of E. faecalis biofilm cells. Moreover, we found that purified recombinant human lysozyme is more effective than CEW lysozyme against E. faecalis biofilms (Fig. 2). While several studies have implicated this enzyme as an antimicrobial capable of reducing biofilm biomass (39–41), it remains to be seen how large of a range of biofilm-forming bacterial species lysozyme could be used to target. Additional promising applications for lysozyme include using it to target polymicrobial biofilms and combining it with clinically relevant antibiotics to enhance disruption of biofilms at infection sites. The data presented here will inform future anti biofilm therapeutic strategies for E. faecalis and other multidrug-resistant pathogens.

MATERIALS AND METHODS

Bacterial strains, growth conditions, and reagents.

E. faecalis strains used in this study are listed in Table 2. All bacterial strains were grown at 37°C under static conditions in brain heart infusion broth (BHI; Becton Dickinson, Franklin Lakes, NJ) or on BHI supplemented with 1.5% agar. E. faecalis biofilms were cultured in tryptic soy broth without added dextrose (TSB_–dex; Becton Dickinson). Micrococcus lysodeikticus, micrococcal nuclease from Staphylococcus aureus, CEW lysozyme, recombinant human lysozyme, ampicillin, and mutanolysin were purchased from Sigma-Aldrich (St. Louis, MO). Restriction enzymes were purchased from New England Biolabs (Ipswich, MA). The human lysozyme-derived peptide LP9 (₁₀₇R-A-W-V-A-W-R-N-R₁₁₅) was synthesized by GenScript (Piscataway, NJ).

Heat inactivation of lysozyme.

Recombinant human lysozyme (1.25 mg/mL in 10 mM Tris-HCl [pH 8]) was boiled at 100°C for 1 h and then placed on ice. Inactivation of lysozyme enzymatic activity was verified by exposing Micrococcus lysodeikticus cells (Sigma-Aldrich) to 250 U/mL of lysozyme, as detailed by the manufacturer. Briefly, 30 μL of heat-treated lysozyme, untreated lysozyme, or reaction buffer (50 mM potassium phosphate) was combined with 800 μL of M. lysodeikticus resuspended in reaction buffer to an OD₄₅₀ of ca. 0.6 to 0.7. The OD₄₅₀ of each solution was measured at 1, 2, 3, 4, 5, and 10 min postincubation.

Strain construction.

Oligonucleotide sequences are listed in Table 3. The OG1RF sigV gene (OG1RF_12448) was deleted in the E. faecalis OG1RF WT and OG1RFΔeep strain backgrounds using the previously described allelic-exchange method (42). The deletion construct used for allelic exchange was generated by overlap-extension PCR with PfuUltra II Fusion HS DNA polymerase (Agilent Technologies, Inc., Santa Clara, CA) to first amplify two ~1-kb fragments from OG1RF genomic DNA with the primer pairs sigV upF/sigV 2stepR and sigV 2stepF/sigV downR. The two products were annealed together, and second step amplification was performed with primers sigV upF and sigV downR. Adenine overhangs were added to the final PCR product with Taq DNA polymerase (New England Biolabs), and the product was ligated into pGEM-T Easy (Promega Corporation, Madison, WI) and then transformed into E. coli DH5α. The resulting plasmid (pGEM-delta sigV) was digested with EcoRI, and the band of interest was gel purified, ligated into EcoRI-digested pCJK47, and propagated in E. coli EC1000.

TABLE 3.

Oligonucleotides used in this study (written 5′ to 3′)

Primer	Sequence (5′–3′)	Source or reference
relA-F	CAAGATTTACGGGTCATTATGG	49
relA-R	GACTAATCCCTAAGCGATGTG	49
dltA RT-F	AACTTTCCAGCATCGCCTGT	This study
dltA RT-R	GGACCAAGTGTTTCCAAAGGC	This study
oatA RT-F	ACCCACACAACTCACACCAG	This study
oatA RT-R	GCCCACTTCCCCATCGATAA	This study
pgdA RT-F	TGTCATGGGGAGGTACAGAA	This study
pgdA RT-R	AGGCTGTCTGTCTAACGGTTC	This study
rsiV RT-F	CCACAAGCCACCACTAAACT	This study
rsiV RT-R	AGAATACCGCCAACAACCGT	This study
sigV RT-F1	GCCTTCAGTGCCTTTTCA	This study
sigV RT-R1	GAAACGCTTGTGGATACC	This study
sigV upF	GAGTCTTCAAGTCCCAGAACC	This study
sigV downR	GACTTGGTTAGGTCAGTAAAA	This study
sigV 2stepR	GCTTCCTTTCTTTTACTAATTCAATAGACTTCGGAGCCAACCATCGAAAACCATCGATTTCTG	This study
sigV 2stepF	CAGAAATCGATGGTTTTCGATGGTTGGCTCCGAAGTCTATTGAATTAGTAAAAGAAAGGAAGC	This study

Viability of planktonic cells on lysozyme-containing agar.

CEW lysozyme solutions were prepared in sterile water and added to molten BHI agar that had been cooled to 55°C. Lysozyme-agar solutions were mixed gently by inversion to avoid the introduction of bubbles and poured into petri dishes to solidify. Overnight BHI broth cultures of OG1RF and OG1RFΔeep were prepared from single colonies. Aliquots of 10-fold serial dilutions prepared from each overnight culture were spread onto BHI agar with and without lysozyme, and CFU/mL values were determined after overnight incubation at 37°C. Three to four biological replicates per lysozyme concentration were performed on separate days.

Lysozyme MIC assay.

The MIC of lysozyme was determined using a broth microdilution method. CEW lysozyme (100 mg/mL dissolved in water) was serially diluted 2-fold in TSB_–dex. Overnight bacterial cultures were diluted 1:100 in TSB_–dex, and then 50 μL of the cell suspension was combined with 50 μL of each lysozyme dilution (0 to 50 mg/mL) in a 96-well U-bottom plate. The plate was incubated at 37°C, and MIC values were assessed after 20 h of growth.

Muramidase and CAMP treatment of planktonic cells in broth culture.

Early logarithmic cultures were prepared by diluting overnight cultures in BHI broth to an OD₆₀₀ of ~0.020. To obtain stationary-phase samples, broth supernatant from overnight bacterial cultures was filtered through a 0.22 μM filter. The spent media were then seeded with cells from an overnight culture to an OD₆₀₀ of ~0.020. CEW lysozyme dissolved in sterile water was added to logarithmic and stationary-phase cultures at a final concentration of 5 mg/mL. Sterile water was added to control samples. Cultures were incubated at 37°C for 6 h. For mutanolysin experiments, cultures were prepared in the same manner as described above and exposed to 500 U/mL of mutanolysin or sterile water. Recombinant human lysozyme and heat-inactivated recombinant human lysozyme (prepared as described above) were prepared in sterile water and added to cultures at a final concentration of 1.25 mg/mL. Polymyxin B was prepared in sterile water and added to cultures at a final concentration of 2 mg/mL. Aliquots of each culture were serially diluted and plated onto BHI agar at 0, 1, 3, and 6 h postexposure to lysozyme. The mean CFU/mL values of three biologically independent experiments are reported.

Microtiter plate biofilm assays.

E. faecalis biofilms were cultured for 20 h in microtiter plates as previously described (19). Microtiter wells were washed three times with sterile water to remove nonadherent cells. Biofilms were treated with lysozyme or ampicillin for 3 h at 37°C under static conditions, unless otherwise indicated. For lysozyme treatment experiments, 100 μL of CEW lysozyme (0.16 to 5 mg/mL in 10 mM Tris-HCl [pH 8]), recombinant human lysozyme (1.25 mg/mL in 10 mM Tris-HCl [pH 8]), 100 μL of heat-inactivated human lysozyme (1.25 mg/mL in 10 mM Tris-HCl [pH 8]), or 10 mM Tris-HCl buffer (pH 8) only was added to each well. For ampicillin treatment experiments, 100 μL of ampicillin (128 μg/mL in water) or water only was added to each well. Exposure of biofilms to other cell envelope-targeting enzymes and CAMPs was accomplished by adding 100 μL of mutanolysin (500 U/mL in water), LP9 (200 μg/mL in 10 mM Tris-HCl [pH 8]), or polymyxin B (200 μg/mL in 10 mM Tris-HCl [pH 8]). Following the specified treatment, biofilms were washed three times with sterile water. To assess cell viability, 100 μL of 10 mM potassium phosphate-buffered saline (KPBS [pH 7.4]) was added to each well (three wells per strain), and then biofilms were scraped with a pipette tip to dislodge bacterial cells. The dislodged cells were serially diluted and plated on BHI agar for enumeration. The cell viability for each biological replicate was calculated as the mean log₁₀ CFU/mL of the three scraped wells. To quantify biofilm biomass, washed wells (up to 13 wells per strain) were allowed to air dry, stained with 0.1% safranin solution for 10 min, washed with water five times, and dried again. Safranin-stained biomass was quantified by measuring the optical density of each well at an absorbance of 450 nm (OD₄₅₀). The OD₄₅₀ values for each plate were averaged. Reported values are the average of three biological replicates. The OD₄₅₀ of safranin-stained wells that contained only sterile growth medium was used for a spectrophotometric blank.

To measure dispersed cells in the biofilm treatment solution supernatant, biofilms were cultivated in microtiter plates and treated with lysozyme or mutanolysin as detailed above. Immediately following treatment, the enzyme or buffer solution was removed from duplicate biofilm-containing wells, serially diluted, and plated onto BHI agar to quantify detached cells. Biofilms were then washed with water three times, and cell viability and biomass of attached cells were measured as described above.

Microscopy.

Biofilms were grown on Aclar fluoropolymer coupons (Electron Microscopy Sciences, Hatfield, PA) for 18 h using previously reported methods (20). Biofilms were washed three times with sterile water and treated with CEW lysozyme (5 mg/mL in 10 mM Tris-HCl buffer) or buffer only for 3 h at 37°C. Following treatment, biofilms were washed with sterile water and labeled using a FilmTracer Live/Dead biofilm viability staining kit (Thermo Fisher Scientific, Waltham, MA) according to the manufacturer’s instructions. In brief, cells were immersed in staining solution consisting of fluorescent nucleic acid dyes SYTO 9 and propidium iodide for 15 min at room temperature. Labeled biofilms were fixed and mounted in ProLong Diamond Antifade Mountant (Thermo Fisher Scientific, Waltham, MA) as detailed by Dale et al. (43). Biofilms were visualized on a Zeiss 700 confocal laser scanning microscope (Carl Zeiss Microscopy, Thornwood, NY) with a 63 × 1.4-numerical-aperture objective housed in the Biomedical Instrumentation Center at Uniformed Services University. Image z-stacks were collected at 0.39-μm intervals, and maximum intensity projections were determined using Zen2012 software (version 14.0.0.201; Carl Zeiss Microscopy). The presented images are representative of two independent biological and three technical replicates per sample.

PicoGreen assay for extracellular DNA detection.

eDNA was detected in biofilms with the Quant-IT PicoGreen dsDNA Reagent (Thermo Fisher Scientific). Biofilms cultured in microtiter plates for 20 h were treated with 100 μL of lysozyme (5 mg/mL in 10 mM Tris-HCl buffer) or buffer only for 3 h at 37°C. Wells containing sterile TSB_–dex were used as blanks. All wells were washed three times with sterile water. Immediately after washing, PicoGreen reagent was added to two wells per sample, and the fluorescence intensity was measured as previously described (44). For nuclease application, lysozyme-treated biofilms were washed, treated with 100 μL of micrococcal nuclease (10 U/mL in reaction buffer) or reaction buffer (1× MNase buffer plus 0.1 mg/mL bovine serum albumin) only for 1 h at 37°C, and then washed again prior to adding PicoGreen reagent. In each experiment, the relative fluorescence units (RFU) were calculated as the average of the biofilm fluorescence – the mean blank fluorescence measured for duplicate wells. After PicoGreen quantitation, biofilms were scraped from two additional wells to quantitate the CFU/mL. All remaining wells were processed as described above to quantitate biofilm biomass via safranin staining. The reported values are the averages of three biological replicates.

Reverse transcription-quantitative PCR.

To prepare biofilms for RNA extraction, overnight E. faecalis cultures were diluted 1:100 into TSB_–dex, and 3 mL of the diluted culture was added to each well of a 6-well plate (Costar 3516; Corning, Inc.). Plates were incubated for 6 h at 37°C with shaking at 120 rpm. Nonadherent planktonic cells from 6 wells were pooled into a single tube, and aliquots were treated with RNAprotect Bacteria Reagent (Qiagen, Hilden, Germany). The adherent biofilms were then gently washed twice with 1 mL of 10 mM KPBS (pH 7.4). Biofilm cells from six wells were dislodged with a cell scraper, pooled into 1 mL of KPBS, and treated with RNAprotect as described previously (19). Planktonic and biofilm cells were collected from three biological replicates on separate days. Total RNA was extracted and converted to cDNA using previously described methods (19, 20). qRT-PCR analyses were performed with Sso Advanced SYBR green Supermix (Bio-Rad Laboratories, Inc., Hercules, CA) in technical triplicate, and threshold cycle (C_T) values were averaged. relA was used as the reference gene for all samples. Primer sequences are listed in Table 3. The fold change of biofilm/planktonic was calculated using the method described by Livak and Schmittgen (45).

Statistical analysis.

Statistical analyses were performed with Prism (version 8.0.0; GraphPad software, Inc., La Jolla, CA). Unless otherwise indicated, two-way analysis of variance (ANOVA) with Sidak’s correction for multiple comparisons was used to assess differences in biofilm biomass and cell viability between treatment and control groups. For experiments in which more than one type of treatment was tested (Fig. 6), a two-way ANOVA with Dunnett’s correction for multiple comparisons was used to identify differences between treatment and control samples.