The Lipopeptide Antibiotic Paenibacterin Binds to the Bacterial Outer Membrane and Exerts Bactericidal Activity through Cytoplasmic Membrane Damage

Abstract

Paenibacterin is a broad-spectrum lipopeptide antimicrobial agent produced by Paenibacillus thiaminolyticus OSY-SE. The compound consists of a cyclic 13-residue peptide and an N-terminal C₁₅ fatty acyl chain. The mechanism of action of paenibacterin against Escherichia coli and Staphylococcus aureus was investigated in this study. The cationic lipopeptide paenibacterin showed a strong affinity for the negatively charged lipopolysaccharides (LPS) from the outer membrane of Gram-negative bacteria. Addition of LPS (100 μg/ml) completely eliminated the antimicrobial activity of paenibacterin against E. coli. The electrostatic interaction between paenibacterin and LPS may have displaced the divalent cations on the LPS network and thus facilitated the uptake of antibiotic into Gram-negative cells. Paenibacterin also damaged the bacterial cytoplasmic membrane, as evidenced by the depolarization of membrane potential and leakage of intracellular potassium ions from cells of E. coli and S. aureus. Therefore, the bactericidal activity of paenibacterin is attributed to disruption of the outer membrane of Gram-negative bacteria and damage of the cytoplasmic membrane of both Gram-negative and Gram-positive bacteria. Despite the evidence of membrane damage, this study does not rule out additional bactericidal mechanisms potentially exerted by paenibacterin.

INTRODUCTION

The extensive spread of antibiotic-resistant bacteria threatens public health worldwide. The Infectious Diseases Society of America recently identified a list of antibiotic-resistant bacterial pathogens that can escape the therapeutic effect of most current antimicrobial agents. These problematic pathogens are Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter spp., which are collectively abbreviated as ESKAPE (1). Thus, there is an urgent need to develop potent and safe antimicrobial agents to combat the emerging antibiotic-resistant pathogens. Paenibacterin is a new cationic lipopeptide antibiotic that is active against Gram-negative and Gram-positive bacteria, including antibiotic-resistant pathogens (2). Paenibacterin consists of a cyclic 13-residue peptide and an N-terminal C₁₅ fatty acyl chain (2). It is produced by a soil isolate, Paenibacillus thiaminolyticus OSY-SE, through nonribosomal peptide synthetases (NRPSs) (3). Aiming for a better understanding of the antimicrobial action, we studied the antibacterial mechanism of paenibacterin against a Gram-negative and a Gram-positive bacterium.

Cationic antimicrobial peptides are positively charged at neutral pH and typically have amphiphilic properties. The cationic and amphiphilic nature of these peptides is important for their antimicrobial activity. Bacterial cell surfaces contain some acidic components, such as lipopolysaccharides (LPS) in Gram-negative bacteria and teichoic acids in Gram-positive bacteria. Cationic antimicrobial peptide can accumulate on the anionic cell surfaces and disrupt the integrity of the cytoplasmic membrane (4). For example, polymyxin is a cationic lipopeptide antibiotic with five positively charged diaminobutyryl (Dab) residues. The hydrophobic domains (N-terminal fatty acyl chain and Phe-Leu segment) and the hydrophilic segments (threonine and Dab residues) of the molecule form two distinct faces, conferring the amphipathicity of polymyxin. Polymyxin exerts its antimicrobial activity against Gram-negative bacteria by permeabilizing the outer membrane and damaging the physical integrity of the phospholipid bilayer of the inner membrane (5). Similarly, the paenibacterin molecule has four positive charges (2). The charged amino acid residues and two polar serine residues constitute the hydrophilic portion of the paenibacterin molecule. Additionally, the N-terminal lipid acyl chain and four aliphatic amino acids (Val6, Ile9, Val11, and Ile13) on one side of the β-sheet structure contribute to the hydrophobicity of paenibacterin (2). The cationic and amphiphilic nature of paenibacterin is assumed to contribute to its broad-spectrum antimicrobial efficacy. The hypothesis of this study is that the cationic paenibacterin binds to anionic cell surfaces and subsequently disrupts the integrity of the cytoplasmic membrane, leading to cell death. To prove the hypothesis, we determined (i) the interaction between paenibacterin and a cell surface component, (ii) the effect of paenibacterin on the cytoplasmic membrane potential, and (iii) the consequences of cytoplasmic membrane damage (i.e., leakage of intracellular contents).

MATERIALS AND METHODS

Effect of lipopolysaccharides (LPS) on paenibacterin activity.

Paenibacterin was prepared and purified as described previously (2). E. coli ATCC 25922 culture, at late exponential phase, was diluted with tryptic soy broth (Becton, Dickinson, Sparks, MD) to contain 6.3 × 10⁷ CFU/ml. Lipopolysaccharides from E. coli O111:B4 (Sigma, St. Louis, MO) were added to the cell suspension at a final concentration of 2, 10, 25, or 100 μg/ml. This was followed by addition of paenibacterin to a final concentration of 32 μg/ml, which was four times the MIC of paenibacterin against E. coli ATCC 25922. Polymyxin E at 2 μg/ml and ampicillin at 8 μg/ml were used as positive and negative controls, respectively. The mixtures were incubated at 37°C with agitation at 200 rpm for 60 min. Survivor cells were quantified by spread-plating on tryptic soy agar.

Binding of paenibacterin to LPS and Gram-negative bacteria.

A fluorescently labeled polymyxin B conjugate (BODIPY FL conjugate; Invitrogen, Carlsbad, CA) was used in this study. This polymyxin B conjugate was used to measure the binding affinity of paenibacterin to purified LPS isolated from E. coli O111:B4. Adding LPS to polymyxin B conjugate quenched its fluorescence due to polymyxin-LPS binding. Aliquots (70 μl) of polymyxin B conjugate-LPS mixture were added to wells of a black nonbinding surface (NBS) microplate (Corning, Tewksbury, MA). This was followed by addition of 30 μl of polymyxin E or paenibacterin at different concentrations. The final concentrations of LPS and polymyxin B conjugate were 5 μg/ml and 0.1 mg/ml, respectively. A change in fluorescence due to the displacement of polymyxin B conjugate from the polymyxin-LPS complex was recorded using a luminescence spectrometer (LS55; Perkin-Elmer, Wellesley, MA), at an excitation wavelength of 490 nm and an emission wavelength of 515 nm. Similarly, the binding affinity of paenibacterin to live E. coli ATCC 25922 cells was measured using the same procedure in which the LPS component was replaced with ∼10⁶ E. coli cells.

Membrane potential depolarization.

The membrane potential of bacterial cells was measured using the fluorescence probe 3,3′-dipropylthiadicarbocyanine iodide [DiSC₃(5); Invitrogen], as described by Zhang et al. (6), with some modifications. The cationic DiSC₃(5) probe can accumulate and integrate into the polarized cell membrane, but depolarization of cell membrane leads to the release of the probe, which changes its fluorescence. An overnight bacterial culture of S. aureus ATCC 6538 or E. coli ATCC 25922 was inoculated (1/100 dilution) into tryptic soy broth and incubated at 37°C with agitation at 200 rpm for 5 h. Bacterial cells were harvested by centrifugation at 3,660 × g at 4°C for 10 min and washed twice using 5 mM HEPES buffer (Sigma) supplemented with 5 mM glucose (buffer A). Cells of S. aureus were resuspended in buffer A. However, E. coli cells were resuspended in a solution composed of buffer A and 0.2 mM EDTA; the latter facilitated the uptake of the DiSC₃(5) probe by this Gram-negative bacterium (6). Subsequently, the fluorescent probe DiSC₃(5) was added to the cell suspensions to a final concentration of 0.5 μM. The mixture was incubated for 15 min at room temperature to allow the uptake of the DiSC₃(5) probe. After incubation, KCl was added to the S. aureus cell suspension at a final concentration of 100 mM, but this step was omitted for E. coli cells. Aliquots (90 μl) of the cell suspension were added to wells of a black NBS microplate. This was followed by the addition of 10 μl of paenibacterin or other antimicrobial agents to the wells. A change in fluorescence due to membrane depolarization was recorded using an LS55 luminescence spectrometer (Perkin-Elmer) at an excitation wavelength of 622 nm and an emission wavelength of 670 nm.

Potassium release assay.

Potassium ions released from bacterial cells were determined using a K⁺-sensitive probe (PBFI; Invitrogen) which is impermeable to bacterial cells (7). Bacterial cells of S. aureus and E. coli were grown, harvested, and washed using the same procedure described above. Aliquots (90 μl) of the cell suspension in buffer A were added to wells of a black NBS microplate. The potassium-sensitive probe, PBFI, was added to the cell suspension at a final concentration of 2 μM. This was followed by the addition of 10 μl of paenibacterin or other antimicrobial agent. A change in fluorescence corresponding to potassium concentration was recorded using a luminescence spectrometer (Perkin-Elmer) at an excitation wavelength of 346 nm and an emission wavelength of 505 nm.

Membrane permeability assay.

Cytoplasmic membrane permeability assays were performed using a vital staining probe mixture (Live/Dead BacLight bacterial viability kit; Invitrogen) according to the manufacturer's instructions. Bacterial cells of S. aureus and E. coli were grown to late exponential phase in tryptic soy broth and washed with saline solution (0.85% NaCl). Bacterial cells suspended in saline were treated with paenibacterin (80 μg/ml for S. aureus; 64 μg/ml for E. coli) at 37°C for 60 min. The two nucleic acid stains provided in the commercial kit, SYTO-9 and propidium iodide (PI), were added to the treated cells at final concentrations of 7.5 μM and 30 μM, respectively. The mixtures were incubated in the dark at 25°C for 15 min. The stained cells (5 μl) were spotted on a microscope slide and covered with a glass coverslip. Digital images were obtained from a fluorescence microscope (BX 61; Olympus, Melville, NY) using the following settings: excitation/emission of 480/500 nm and 490/635 nm for SYTO-9 and propidium iodide, respectively.

Statistical analysis.

Experiments to measure fluorescence and bacterial population counts were repeated independently at least two times. For fluorescence measurements, the time points where the curves became asymptotic were subjected to statistical analyses. For bacterial inactivation assays, the survivor cell counts at the terminus of experiments were analyzed. All data were subjected to analysis of variance followed by Tukey's honest significant difference (HSD) tests using SPSS (version 20; SPSS, Inc., Chicago, IL, USA).

RESULTS

Lipopolysaccharides antagonize paenibacterin activity.

Purified lipopolysaccharides (LPS) from E. coli affected the antimicrobial activity of paenibacterin against E. coli ATCC 25922. At low concentrations (2 to 25 μg/ml), LPS had little effect on the antimicrobial activity of paenibacterin; however, at 100 μg/ml, LPS completely neutralized the bactericidal activity of paenibacterin (Fig. 1). Polymyxin E, a known outer membrane binding agent which was used as a positive control, became inactive against the bacterium in the presence of 100 μg of LPS/ml. The negative control, ampicillin, which targets cell wall biosynthesis, was not affected by LPS (Fig. 1). The result suggests that paenibacterin has a high affinity to the LPS component from the outer membrane of Gram-negative bacteria. Therefore, the LPS component in live bacterial cells is likely the initial target of paenibacterin.

FIG 1.

Escherichia coli ATCC 25922 survival after treatment with paenibacterin (32 μg/ml), polymyxin E (2 μg/ml), or ampicillin (8 μg/ml) without or with lipopolysaccharide (LPS) at different concentrations. Values are expressed as means (number of experiments, 2), and error bars represent standard deviations. Means with different letters are significantly different (P < 0.05).

Paenibacterin binds to the LPS component of the outer cell membrane.

Preliminary data show that the fluorescence of polymyxin B conjugate was quenched when this probe bound to a purified LPS component or live E. coli cells. Compounds with affinity to LPS may compete with polymyxin B conjugate for the binding sites on LPS and thus displace the conjugate from LPS, with a concomitant increase in fluorescence. Addition of polymyxin E to the polymyxin B conjugate-LPS complex caused an increase in fluorescence, which indicated that the fluorescently labeled polymyxin B molecules were displaced and released from LPS (Fig. 2A). The result is consistent with previous reports that cationic polymyxin molecules can bind to the negatively charged lipid A phosphates in LPS and displace the divalent cations (Mg²⁺ and Ca²⁺) which stabilize the permeability barrier of the outer membrane (5, 8). Similarly, cationic paenibacterin molecules competed with polymyxin B conjugate for binding on LPS, resulting in the release of the conjugate in a concentration-dependent manner (Fig. 2A). Moreover, paenibacterin can also displace the membrane-bound polymyxin molecules from live E. coli cells (Fig. 2B). These results suggest that paenibacterin and polymyxin molecules have the same binding sites on LPS in the outer membrane of Gram-negative bacteria. Therefore, the electrostatic interaction between paenibacterin and LPS may disrupt outer membrane permeability and promote the uptake of the antibiotic.

FIG 2.

Displacement of fluorescence-labeled polymyxin B conjugate from lipopolysaccharide (LPS) reagent or live Escherichia coli ATCC 25922 cells by different concentrations of paenibacterin. (A) In vitro polymyxin B-LPS displacement. (B) Polymyxin B cell displacement assay. Values are expressed as means (n = 3 for in vitro displacement assays; n = 2 for cell displacement assays), and error bars represent standard deviations. Asterisks (*) indicate a statistically significant difference (P < 0.05) between the control and the treatments. paen, paenibacterin; pmx, polymyxin E.

Paenibacterin depolarizes cytoplasmic membrane.

Aerobic or facultative bacteria maintain a proton gradient (proton motive force) across the cell membrane that can be used to generate ATP and transport nutrients into the cells. The main component of the proton motive force is an electrical potential gradient (Δψ) (9). DiSC₃(5) probe, a membrane potential-sensitive dye, accumulates in healthy polarized cell membrane and becomes self-quenched. As shown in Fig. 3A, paenibacterin at a concentration of 32 μg/ml significantly permeabilized the cell membrane of S. aureus. Compounds which depolarize the membrane electrical potential lead to a release of DiSC₃(5) from cell membrane, resulting in an increase in fluorescence (6). The positive control, nisin, is a membrane-active peptide. Treatment with nisin (25 μg/ml) also caused considerable membrane depolarization. The negative control, vancomycin, is a cell wall inhibitor. The presence of vancomycin had little effect on cell membrane charge. For E. coli, a Gram-negative bacterium, paenibacterin at 32 μg/ml caused a significant membrane potential disturbance; the antibiotic at 64 and 128 μg/ml induced much higher levels of membrane depolarization in this bacterium (Fig. 3B).

FIG 3.

Changes in bacterial membrane potential in the presence of paenibacterin, nisin, vancomycin, and polymyxin E. (A) Staphylococcus aureus ATCC 6538. (B) Escherichia coli ATCC 25922. Values are expressed as means (n = 2), and error bars represent standard deviations. Asterisks (*) indicate a statistically significant difference (P < 0.05) between the control and treatments. paen, paenibacterin; pmx, polymyxin E; van, vancomycin.

Paenibacterin releases intracellular potassium ions.

Membrane depolarization by paenibacterin prompted us to examine whether other membrane-associated functions were affected. Paenibacterin treatment caused a concentration-dependent K⁺ leakage from cells of S. aureus and E. coli (Fig. 4). At low concentrations (i.e., 4× MIC), paenibacterin did not lead to significant potassium ion leakage from cells of S. aureus or E. coli. Similar results were observed for nisin at 25 μg/ml and polymyxin E at 10 μg/ml. At higher concentrations (64 and 128 μg/ml), paenibacterin rapidly induced the release of intracellular potassium ions (Fig. 4). These results provide additional evidence that paenibacterin compromised the integrity of bacterial cytoplasmic membrane.

FIG 4.

Release of intracellular K⁺ in the presence of paenibacterin, nisin, vancomycin, and polymyxin E. (A) Staphylococcus aureus ATCC 6538. (B) Escherichia coli ATCC 25922. Values are expressed as means (n = 3 for S. aureus; n = 2 for E. coli), and error bars represent standard deviations. Asterisks (*) indicate a statistically significant difference (P < 0.05) between the control and the treatments. paen, paenibacterin; pmx, polymyxin E; van, vancomycin.

Paenibacterin increases cell membrane permeability.

The permeability of the cytoplasmic membrane was visualized after staining with two fluorescent nucleic acid stains, SYTO-9 and propidium iodide (PI). Viable cells with intact cytoplasmic membrane appear green when the membrane-permeable stain, SYTO-9, enters the cells. Bacterial cells with compromised membranes are stained red by propidium iodide, a membrane-impermeant stain. The untreated S. aureus and E. coli cells were stained green, whereas the majority of bacterial cells treated by paenibacterin for 60 min became red (Fig. 5). These results clearly show increased permeability of the cytoplasmic membrane after Gram-positive and Gram-negative cells were treated with paenibacterin.

FIG 5.

Changes in bacterial cell membrane permeability, as observed under a fluorescence microscope, when cells were treated with paenibacterin and tested with a viable-cell staining technique; membrane is expected to be intact in green fluorescent and compromised in red fluorescent cells. (A) Untreated cells of Staphylococcus aureus ATCC 6538. (B) S. aureus ATCC 6538 treated with 80 μg/ml paenibacterin for 60 min. (C) Untreated Escherichia coli ATCC 25922. (D) E. coli ATCC 25922 treated with 64 μg/ml paenibacterin for 60 min.

DISCUSSION

This study explored the mechanism of action of paenibacterin against both Gram-positive and Gram-negative bacteria. Gram-negative bacteria are naturally resistant to many antibiotics due to the outer membrane barrier (10). The molecular basis of permeability lies in the lipopolysaccharide networks that are electrostatically linked by divalent cations (Mg²⁺ and Ca²⁺). Cationic agents and ion chelators are well-known permeabilizers that can disorganize the outer membrane (10). Cationic peptides, such as polymyxins, enter bacterial cells by a self-promoted uptake pathway (8). In this study, we found that paenibacterin had a high affinity for LPS (Fig. 1 and 2). The interaction between paenibacterin and LPS may displace Mg²⁺ and Ca²⁺ and thus promote the uptake of paenibacterin. The LPS binding ability of paenibacterin suggested that paenibacterin may neutralize LPS and minimize endotoxemia during antibiotic treatment (8). Moreover, as an outer membrane permeabilizer, paenibacterin combined with other antibiotics may have a synergistic effect against drug-resistant pathogens.

The bacterial cytoplasmic membrane is the target of paenibacterin in both Gram-positive and Gram-negative bacteria. Paenibacterin depolarizes the cell membrane, triggers K⁺ release, and increases the uptake of the membrane-impermeant probe propidium iodide (Fig. 3, 4, and 5). The bacterial cytoplasmic membrane is essential to bacterial life as it contains one-third of the proteins in a cell and is the place for many crucial processes (11). Therefore, paenibacterin may interfere with numerous cellular functions by damaging the cell membrane. Unlike other conventional antibiotics, membrane-active peptides are less likely to mediate antibiotic resistance in treated bacteria (4, 11). In terms of the challenges of antibiotic resistance, paenibacterin is a potential candidate for treatment of emerging infectious diseases. Other lipopeptides, such as tripropeptin C (12) and friulimicin B (13), inhibit cell wall synthesis by forming a complex with the lipid carriers which deliver the precursors for cell wall biosynthesis. This mechanism of action has not been investigated in the current study. Although paenibacterin clearly damages cytoplasmic membranes, this study does not rule out additional bactericidal mechanisms.

To study the LPS-paenibacterin interaction, it was necessary to develop a rapid method for determining the binding affinity of paenibacterin and other compounds to LPS. We used a commercial polymyxin B conjugate (BODIPY FL conjugate-polymyxin B) to perform these assays. The fluorescence of the labeled polymyxin becomes quenched when the probe binds to purified LPS or Gram-negative cells. Other compounds with affinity to LPS can release the labeled polymyxins, which leads to an increase in fluorescence. This approach is a convenient alternative to methods relying on dansyl-polymyxin reagent to study the interaction between polycations and LPS (14). The new method can be extended to screen for compounds with antiendotoxin activity.