Antibacterial properties and mode of action of a short acyl-lysyl oligomer

Abstract

We investigated the potency, selectivity, and mode of action of the oligo-acyl-lysine (OAK) NC(12)-2 beta(12), which was recently suggested to represent the shortest OAK sequence that retains nonhemolytic antibacterial properties. A growth inhibition assay against a panel of 48 bacterial strains confirmed that NC(12)-2 beta(12) exerted potent activity against gram-positive bacteria while exhibiting negligible hemolysis up to at least 100 times the MIC. Interestingly, NC(12)-2 beta(12) demonstrated a bacteriostatic mode of action, unlike previously described OAKs that were bactericidal and essentially active against gram-negative bacteria only. The results of various experiments with binding to model phospholipid membranes correlated well with those of the cytotoxicity experiments and provided a plausible explanation for the observed activity profile. Thus, surface plasmon resonance experiments performed with model bilayers revealed high binding affinity to a membrane composition that mimicked the plasma membrane of staphylococci (global affinity constant [K(app)], 3.7 x 10(6) M(-1)) and significantly lower affinities to mimics of Escherichia coli or red blood cell cytoplasmic membranes. Additional insertion isotherms and epifluorescence microscopy experiments performed with model Langmuir monolayers mimicking the outer leaflet of plasma membranes demonstrated the preferential insertion of NC(12)-2 beta(12) into highly anionic membranes. Finally, we provide mechanistic studies in support of the view that the bacteriostatic effect resulted from a relatively slow process of plasma membrane permeabilization involving discrete leakage of small solutes, such as intracellular ATP. Collectively, the data point to short OAKs as a potential source for new antibacterial compounds that can selectively affect the growth of gram-positive bacteria while circumventing potential adverse effects linked to lytic compounds.

FIG. 1.

Biophysical characteristics of NC₁₂-2β₁₂. (a) OAK molecular structure (mass, 1,121.7 g/mol). (b and c) Effects of ionic strength and pH on antimicrobial activity, as assessed against two strains of S. aureus: ATCC 29213 (squares) and ATCC 25923 (triangles). (d) Dose-dependent hemolytic activity of NC₁₂-2β₁₂ (stars) and C₁₆-K-β₁₂ (circles) as determined by measuring the absorbances of the supernatants (at 470 nm) after 1 h of incubation with RBCs (10%; 37°C) in PBS. (e) Dose-dependent light-scattering intensities of PBS solutions of NC₁₂-2β₁₂ (stars) and C₁₆-K-β₁₂ (circles). The error bars indicate standard deviations calculated from at least three independent experiments performed in duplicate. Lack of standard deviations indicates consistency.

FIG. 2.

Mechanistic studies. (a) Ability of a representative S. aureus strain to form colonies upon exposure to the OAK. Open circles, control; open stars, NC₁₂-2β₁₂ at 14 μg/ml; closed stars, NC₁₂-2β₁₂ at 28 μg/ml. Bacteria were sampled at the indicated times, subjected to serial 10-fold dilutions, and plated on LB agar dishes for CFU counts after overnight incubation at 37°C. (b and c) Extracellular (b) and total (c) ATP levels in treated bacterial suspensions. Total ATP was measured after extraction with DMSO. Open circles, control; closed stars, NC₁₂-2β₁₂ (25 μM); open triangles, dermaseptin S4(1-16) (25 μM). RLU, relative light units. (d) EtBr accumulation in S. aureus ATCC 29213 suspended in PBS. EtBr was added (1 μg/ml) at time zero, and the fluorescence change over time was recorded. OAK concentrations were 14, 7, and 3.5 μg/ml, respectively, for the dark, light, and dotted lines. The circles represent control experiments: black circles, Triton (0.01%); open circles, EtBr alone or in the presence of gentamicin, ciprofloxacin, and tetracycline (four times the MIC each). Note that all lines were from continuous recordings; the symbols were added only for clarity. The error bars indicate standard deviations.

FIG. 3.

Properties of binding to model membranes. Shown are overlays of typical sensorgrams for various concentrations of NC₁₂-2β₁₂ in PBS, using three different model membranes: POPG-CL (1.4/1) (a); POPG-PE (1/4) (b); cholesterol-PC (1/9) (c). In each panel, the curves with higher intensities correspond to higher OAK concentrations. The resulting affinity constants determined by numerical integration using the two-step reaction model (17) are listed in Table 2.

FIG. 4.

Properties of binding to Langmuir monolayers. (a) Insertion isotherms of DPPG, DPPG-DPPE (1/4), and cholesterol-DPPC (1/9) monolayers showing the percentage of change in area per molecule after injection of NC₁₂-2β₁₂ (5.6 μg/ml). (b) Epifluorescence images of DPPG and DPPG-DPPE (1/4) before and after NC₁₂-2β₁₂ injection.

FIG. 5.

DNA binding and resistance studies. (Top) Representative agarose gel runs of the bacterial plasmid pUC19 (150 ng) in its native state and after exposure to BamHII (lanes 2 and 3, respectively). Lanes 4 and 5, pUC19 after preincubation with NC₁₂-2β₁₂ (14 and 28 μg/ml, respectively), followed by exposure to BamHI; lane 6, pUC19 after preincubation with C₁₂-K7α₈ (21 μg/ml), followed by exposure to BamHI; lane 1, molecular weight marker. (Bottom) Evolution of MICs after successive exposures of S. aureus (ATCC 29213) to subinhibitory concentrations of the specified antimicrobial agent: oxacillin (circles), ciprofloxacin (triangles), or NC₁₂-2β₁₂ (open stars). Plotted is the MIC evolution for 15 serial passages. The relative MIC is the normalized ratio of the MIC obtained for a given subculture to the MIC obtained for first-time exposure (determined concomitantly).