Molecular basis for the increased polymyxin susceptibility of Klebsiella pneumoniae strains with under-acylated lipid A

Abstract

The impact of under-acylation of lipid A on the interaction between Klebsiella pneumoniae LPS and polymyxins B and E was examined with fluorometric and calorimetric methods, and by ¹H NMR, using a paired wild type (WT) and the ΔlpxM mutant strains B5055 and B5055ΔlpxM, which predominantly express LPS with hexa- and penta-acylated lipid A structures respectively. LPS from B5055ΔlpxM displayed a fourfold increased binding affinity for polymyxins B and E compared with the B5055 WT LPS. EC₅₀ values were consistent with polymyxin minimum inhibitory concentration (MIC) values for each strain. Accordingly, polymyxin exposure considerably enhanced the permeability of the B5055ΔlpxM OM. Analysis of the melting profiles of isolated LPS aggregates suggested that bactericidal polymyxin activity may relate to the acyl chains’ phase of the outer membrane (OM). The enhanced polymyxin susceptibility of B5055ΔlpxM may be attributable to the favorable insertion of polymyxins into the more fluid OM compared with B5055. Molecular models of the polymyxin B–lipid A complex illuminate the key role of the lipid A acyl chains for complexation of polymyxin. The data provide important insight into the molecular basis for the increased polymyxin susceptibility of K. pneumoniae strains with under-acylated lipid A. Under-acylation appears to facilitate the integration of the N-terminal fatty-acyl chain of polymyxin into the OM resulting in an increased susceptibility to its antimicrobial activity/activities.

Keywords

Polymyxin lipopolysaccharide

Introduction

Klebsiella pneumoniae is an opportunistic pathogen, which has been frequently implicated in serious nosocomial infections¹ and associated with high rates of mortality,² particularly in immuno-compromised patients. Concerns have thus been generated in response to the escalation in the incidence of multidrug-resistant (MDR) K. pneumoniae infections,^3–5 with the realization that effective therapeutic options are becoming increasingly limited. In view of the waning antibacterial development pipeline, interest in polymyxin E (syn. colistin) and polymyxin B has been renewed and both are increasingly used as a last-line therapy.⁶ Indeed, considerable in vitro activity against K. pneumoniae strains has been demonstrated; 98.2% of general clinical isolates of K. pneumoniae were susceptible to polymyxins.⁷ Extremely drug-resistant strains, which are also resistant to polymyxins, have emerged.^8,9 These findings demand a greater appreciation of the mechanism(s) of polymyxin activity and resistance in K. pneumoniae in order to minimize the development of resistance and assist drug discovery strategies.

The Gram-negative outer membrane (OM) is a permeability barrier to various antimicrobial substances, including numerous antibacterials.^10,11 This complex asymmetrical structure comprises an inner phospholipid leaflet, as well as an outer leaflet, which predominantly contains LPS. Structurally, LPS is composed of three domains: the highly variable and serovar-dependent O-antigen chain (encompassing repeated saccharide units) is linked to a more genus-related core oligosaccharide region and these extra-cellular sugar portions are built upon the relatively conserved lipid A moiety (Figure 1). Lipid A is intercalated within the membrane functioning as a hydrophobic anchor.

Figure 1.

Silver stained SDS-PAGE analysis of purified LPS from the Klebsiella pneumoniae B5055 strains examined in this study. Increasing sample amounts (0.5, 1, 1.5, 2 µg) were loaded per lane to allow for visualization owing to the differential staining of each sub-structure. The position of each LPS sub-structural component is indicated on the right ordinate. Molecular mass standards are shown on the left ordinate, in lane 1 of the gel. The dashed line indicates the mass shift of the lipid A between the paired strains. The structural organization of LPS is shown schematically.

For polymyxins, lipid A represents an important binding target.¹² The cationic amphipathic nature of polymyxins is understood to be crucial to enable ‘self-promoted uptake’ across the OM.¹³ The importance of hydrophobic interactions with the lipid A acyl chains for the antibacterial activity of polymyxins against K. pneumoniae has been highlighted in a recent study, which showed that K. pneumoniae strains that express an LPS chemotype with an under-acylated lipid A displayed an increased susceptibility to polymyxins.¹⁴ The relationship between polymyxin susceptibility and lipid A structure has been poorly characterized in K. pneumoniae and warrants further investigation. In the present study we employed a series of biophysical methods to examine the correlation between the increased polymyxin susceptibility of K. pneumoniae and the under-acylated lipid A; to this end, we employed a mutant strain B5055ΔlpxM,¹⁴ harboring a defective lpxM gene (formally msbB or waaN) that encodes the enzyme responsible for the late secondary acylation of the immature lipid A structure. This modification results in the generation of a predominantly penta-acylated lipid A as opposed to the parent B5055 wild type (WT) strain, which predominantly expresses a hexa-acylated lipid A.¹⁴ Furthermore, the effect of polymyxin exposure on the phase transition temperature of the LPS hydrocarbon moieties of each test strain was investigated by fluorometric thermal shift analysis. Finally, on the basis of these data structural correlations are drawn using molecular models of the polymyxin B₁–K. pneumoniae lipid A complexes.

Materials and methods

Materials

Polymyxin B (sulfate), polymyxin E (sulfate), 1-N-phenylnaphthylamine (NPN), lysozyme, purpald reagent, 3-deoxy-D-manno-oct-2-ulopyranosonic acid (Kdo) and deoxycholate were obtained from Sigma-Aldrich (Sydney, NSW, Australia). All other reagents were of the highest purity commercially available.

Bacterial strains

Klebsiella pneumoniae B5055 is a mouse virulent clinical isolate (serotype K2 : O1) and B5055ΔlpxM is a mutant derivative constructed with a defective lpxM gene.¹⁴ The lipid A structure of both strains has been characterized comprehensively in a previous study.¹⁴

Determination of minimum inhibitory concentrations

Polymyxin minimum inhibitory concentrations (MICs) for each strain were determined by the broth microdilution method.¹⁵ Experiments were performed with cation-adjusted Mueller-Hinton broth (CaMHB) (containing approximately 10⁶ CFU per ml) in 96-well polypropylene microtitre plates. Wells were inoculated with 100 µl of bacterial suspension prepared in CaMHB and 100 µl of CaMHB containing increasing concentrations of polymyxins (0 to 128 µg/ml) over 7 titre steps. The MICs were defined as the lowest concentration at which visible growth was inhibited following 18 h incubation at 37℃. Cell viability was determined by sampling wells at polymyxin concentrations greater than the MIC. These samples were diluted in normal saline and spread plated onto nutrient agar. After incubation at 37℃ for 20 h, viable colonies were counted on these plates. The limit of detection was 10 CFU/ml. European Committee on Antimicrobial Susceptibility Testing (EUCAST; Växjö, Sweden) breakpoints against Enterobacteriaceae were employed.¹⁶

LPS isolation

LPS was extracted by a modified hot phenol/water procedure.¹⁷ The purity of isolated LPS samples was ascertained by SDS-PAGE. LPS samples were re-suspended in SDS-PAGE sample buffer (2.3% SDS, 0.8% Tris, 10% glycerol, 5% dithiothreitol) and boiled for 10 min. SDS-PAGE in 15% polyacrylamide gels and LPS was visualized by silver staining.¹⁸

Kdo purpald assay

The molar concentration of Kdo in the LPS samples was measured following the purpald assay, as described in detail previously.¹⁹ The un-substituted terminal vicinal glycol group of Kdo in LPS is subjected to sodium periodate (NaIO₄) oxidation, yielding quantitative formaldehyde, which is measurable by purpald reagent. Briefly, 50 µl of 32 mM NaIO₄ was added to 50 µl of LPS sample or Kdo standard (0.05, 0.1, 0.2, 0.3, 0.4 and 0.5 mM Kdo) and incubated for 25 min. Fifty microliters of 136 mM purpald reagent in 2 N NaOH was added to each sample and incubated for 20 min. Then, 50 µl of 64 mM NaIO₄ was added to the samples and incubated for another 20 min. The absorbance of each sample was measured at 550 nm. The Kdo molarity in the LPS samples was evaluated from their absorbance and the Kdo standard curve. This value divided by the theoretical number of Kdo residues per LPS molecule gave the molarity of LPS in the sample.

Fluorometric measurement of [Dansyl-Lys]¹- polymyxin B₃ binding to whole cells and isolated LPS

The fluorometric measurements were performed as described previously using the fully synthetic antibacterial polymyxin probe DPmB₃.²⁰

Deoxycholate sensitivity induced by polymyxins

The protocol for this assay was adapted from Vaara et al.²¹ with minor modifications. Mid-logarithmic phase cells (OD_500nm = 0.4 to 0.6) were pelleted and re-suspended in 5 ml of 5 mM HEPES buffer (pH 7.2) in the presence of increasing concentrations of polymyxin E, then incubated for 10 min. After the OD_500nm of the cell suspensions was measured, cells were again pelleted and re-suspended in 5 ml of the same buffer containing 0.25% (w/v) deoxycholate. These suspensions were incubated for 10 min and the decrease in OD_500nm was subsequently measured. All assay operations were performed at 37℃.

Permeability to lysozyme induced by polymyxins

This method was performed as described previously.²¹ Lysozyme (5 µg/ml) and polymyxin E (9 µg/ml) were added to a mid-logarithmic phase cell suspension (OD_500nm = 0.4 to 0.6), and the decrease in OD_500nm was measured.

NPN uptake assay

NPN uptake was measured as described previoulsy.²² Briefly, mid-logarithmic phase cells were pelleted and washed twice in 5 mM HEPES buffer (pH 7.2) containing 1 mM sodium azide, then re-suspended in the same buffer (OD_500nm = 0.5). The cell suspension was allowed to sit at room temperature for 30 min before analysis. NPN fluorescence was measured with the excitation and emission wavelengths set at 350 and 420 nm, respectively, with slit widths of 5 nm. The NPN uptake assay was also employed to measure the OM permeabilizing activity of polymyxins. NPN was dissolved in acetone at a concentration of 500 µM and added to 1 ml of cell suspension to a final concentration of 10 µM. Response-concentration curves were plotted from baseline-corrected values and non-linear regression was used to fit equation (1) to the data

Δ F = F_{min} + \frac{F_{max} - F_{min}}{1 + 10^{(\log {EC}_{50} - [polymyxin]) \times Hill Slope}}

(1)

Where ΔF represents the NPN fluorescence enhancement upon the addition of polymyxin B or E to the cell suspension; F_min is the minimum NPN fluorescence enhancement produced by polymyxin treatment and is defined by the bottom plateau of the dose-response curve; F_max is the maximum fluorescence enhancement upon the addition of polymyxin; [polymyxin] represents the free concentration of polymyxins; and EC₅₀ represents the concentration of polymyxins required to produce half F_max determined from the data fit.

Fluorometric thermal shift assay

The phase transition of the fatty acyl chains of the LPS aggregates was examined in the absence and presence of polymyxin E via a NPN fluorometric thermal shift assay. Fluorescence measurements were performed on a Cary Eclipse spectrophotofluorometer, as detailed earlier. LPS (50 µM) dissolved in 20 mM HEPES buffer (pH 7.4) together with 5 µM NPN was heated from a starting temperature of 5℃ to a target temperature of 95℃ in 1℃ increments. Data-fitting operations were performed as described previously to obtain information on the phase transition range and its mid-point (T_m).²³ The molar concentration of LPS was determined using the purpald Kdo assay.¹⁹

Isothermal titration calorimetry measurement of polymyxin–LPS binding interactions

Microcalorimetric experiments of polymyxin E binding to LPS were performed on a VP-ITC isothermal titration calorimeter (Microcal, Northampton, MA, USA) at 37℃ in 20 mM HEPES buffer (pH 7.0), as described previously.²⁰ The molar concentration of LPS was determined using the purpald Kdo assay.¹⁹

¹H NMR spectroscopy measurements

Polymyxin E was dissolved in 0.5 ml of ²H₂O (pH 4.0) to a final concentration of 0.1 mM. In order to examine the binding of polymyxin E to LPS line-broadening experiments were performed by successively adding small aliquots of a solution of LPS to polymyxin E. One-dimensional proton NMR (¹H NMR) spectra were acquired on a Varian INOVA 600 MHz spectrometer (Varian, Mulgrave, VIC, Australia) operating at 20℃. The chemical shifts (ppm) were referenced to an internal 3-(trimethylsilyl)-1-propanesulfonic acid sodium salt (TMS).

Molecular modeling of the lipid A–polymyxin B complex

The coordinates of the NMR solution structure of polymyxin B₁ when bound to Escherichia coli LPS were obtained from Pristovsek and Kidric.^24,25 The lipid A structures of K. pneumoniae B5055 WT and B5055 ΔlpxM¹⁴ were modeled using the crystallographic structure of E. coli LPS as a template; the latter was derived from the crystal structure of FhuA, the receptor for ferrichrome-iron in E. coli with bound LPS (PDB code:1QFF).^12,26 The polymyxin B₁–lipid A complexes were manually docked to reproduce the interactions described by Pristovsek and Kidric^24,25 using Accelrys Discovery Studio V2.1. (Accelrys, Inc. San Diego, CA, USA).

Results

Susceptibility testing and LPS characterization

Polymyxin MICs are presented in Table 1 for each strain. K. pneumoniae B5055 ΔlpxM was more susceptible to polymyxins (MIC = 0.25 µg/ml and 0.125 µg/ml, polymyxin B and E respectively) compared with strain B5055 (MIC = 2 µg/ml and 4 µg/ml, polymyxin B and E respectively; Table 1). SDS-PAGE analysis of purified LPS samples revealed noticeable differences in the size of the lipid A component from each strain, as reflected by the extent of lipid A migration through the gel (Figure 1). The lipid A of B5055 ΔlpxM migrated closer to the gel front compared with that of its parent strain B5055, which is consistent with the loss of secondary acylation in the lipid A component and lower mass. A ladder-like pattern was observed for the LPS isolated from the paired B5055 strains, representative of the repeating sugar units in the O-antigen chain.²⁷

Table 1.

K. pneumoniae susceptibility to polymyxin B and E, and half-maximal effective concentration (EC₅₀) of polymyxins required to produce permealization of the OM to NPN.

Strain	Polymyxin B		Polymyxin E
Strain	MIC (mg/l)	NPN EC₅₀ (µM)	MIC (mg/l)	NPN EC₅₀ (µM)
B5055	2	0.48 ± 0.15	4	0.49 ± 0.08
B5055ΔlpxM	0.125	0.19 ± 0.05	0.25	0.15 ± 0.06

MIC, minimum inhibitory concentration.

Binding of polymyxin B and E to isolated K. pneumoniae LPS

Fluorometric assessment of DPmB₃ binding to isolated LPS

The binding affinity of polymyxin for purified LPS from each strain was evaluated using a synthetic fluorescent polymyxin probe, DPmB₃,²⁰ from which the affinity (K_d) was calculated (Figure 2B; Table 2).²⁰ Subsequently, the decrease in fluorescence upon displacement of this probe by unlabeled polymyxins was analyzed to determine the binding affinity data (K_i) for polymyxins B and E (Figure 2C; Table 2). The LPS binding affinity constants determined for polymyxins B and E were comparable with the affinity constants obtained for DPmB₃ (Table 2). All three compounds were shown to bind with greater affinity to the LPS from strain B5055 ΔlpxM, in comparison with the LPS from the B5055 WT strain (Figure 2B, C; Table 2).

Table 2.

Binding affinity of polymyxins for purified LPS from each K. pneumoniae test strain.

K. pneumoniae LPS	K_d^a; K_i^b Binding affinity (µM)
K. pneumoniae LPS	¹[Dansyl]PmB₃^a	Polymyxin B^b	Polymyxin E^b
K. pneumoniae B5055	0.67 ± 0.21	0.73 ± 0.14	0.87 ± 0.21
K. pneumoniae B5055ΔlpxM	0.33 ± 0.08	0.15 ± 0.03	0.18 ± 0.05

K_d, dissociation constant value, a direct measure of the ¹[Dansyl]PmB₃ binding affinity.

K_i, inhibition constant value, measured as the ability of the polymyxin to displace ¹[Dansyl]PmB₃.

Figure 2.

(A) Fluorescence emission spectra of (i) K. pneumoniae B5055 LPS in buffer (3 mg/l); (ii) [Dansyl-Lys]¹-polymyxin B₃ (12 µM) in buffer; and (iii) [Dansyl-Lys]¹-polymyxin B₃ (12 µM) bound to B5055 LPS (3 mg/l). (B) Fluorescence assay of the binding of [Dansyl-Lys]¹-polymyxin B₃ to isolated LPS from K. pneumoniae B5055 (♦) and B5055ΔlpxM (○).The solid lines represent the non-linear least squares regression fit of a one-site binding model as described previously,²⁰ R² values for the fits ranged between 0.93 and 0.98. (C) The decrease in fluorescence upon displacement of [Dansyl-Lys]¹-polymyxin B₃ from LPS K. pneumoniae B5055 (♦) and B5055ΔlpxM (○) by titration with polymyxin E. Solid lines represent non-linear least squares regression fit of a displacement model as detailed previously.²⁰ All data points are the mean of three independent measurements ± SD.

¹H NMR analysis of the binding of polymyxin E with LPS aggregates

Titration of LPS aggregates into a solution of polymyxin E produced a line broadening of all polymyxin resonances; for simplicity, only the amide region is shown in Figure 3. Based on an identical titration regime mass ratio (polymyxin E:LPS), line broadening was observed at a lower mass ratio for titrations with B5055 ΔlpxM LPS compared with LPS from the B5055 WT strain. This suggests that polymyxin E penetrates penta-acylated B5055 ΔlpxM LPS aggregates more readily than it does the hexa-acylated B5055 LPS phenotype (Figure 3A, B).

Figure 3.

Six hundred MHz ¹H NMR spectrum of the amide region of polymyxin E in 10% D₂O pH 4.0 titrated with LPS. The ratio of LPS to polymyxin E is indicated on the right ordinate. (A) KP B5055 LPS (B) KP B5055ΔlpxM LPS.

Isothermal titration calorimetry assay of polymyxin E binding to isolated LPS

The energetics of the interaction between polymyxin E and LPS from the paired B5055 strains was determined by isothermal titration calorimetry (ITC). The microscopic thermodynamic parameters, the number of binding sites (n), the binding constant (K_a) and binding enthalphy (ΔH) are documented in graphical form as insets in Figure 4. As the nature of the polymyxin B–LPS interaction has been reported to be temperature dependent,²⁸ all ITC experiments were performed at a physiologically relevant temperature of 37℃. The downward peaks observed for polymyxin E titrations with both B5055 and B5055 ΔlpxM LPS were representative of an exothermic process (Figure 4, top panels). Plots of the enthalpy changes measured calorimetrically versus the polymyxin E:LPS ratio are also presented in Figure 4 (bottom panels). The data were evaluated according to a model which defines two independent binding sites, as described previously.²⁰ The microscopic thermodynamic parameters for the first set of sites for the polymyxin E–B5055 LPS (hexa-acylated lipid A) interaction revealed an approximately 0.5 : 1 (polymyxin E:LPS) binding stoichiometry, with a moderate binding affinity (6.8 µM). This reaction displayed an unfavorable enthalpic component and a large positive (favorable) entropic component (Figure 4A, insets). The parameters for the second set of sites indicated a 4 : 1 (polymyxin E:LPS) binding stoichiometry with a moderate binding affinity (3.1 µM) and was driven by favorable enthalpic and entropic contributions to binding free energy (Figure 4A, insets). In contrast, the microscopic thermodynamic parameters for the polymyxin E–B5055 ΔlpxM LPS clearly indicated a more thermodynamically favorable binding process; the first set of sites of the interaction revealed a 1 : 1 (polymyxin E:LPS) stoichiometry with a high binding affinity (0.06 µM) (Figure 4B, insets). The reaction was largely driven enthalpically and accompanied by a smaller, albeit favorable, entropic contribution to the binding free energy (Figure 4B, insets). The second phase of the polymyxin E–B5055 ΔlpxM LPS reaction was characterized by a 4 : 1 (polymyxin E:LPS) binding stoichiometry with a moderate binding affinity (13 µM) (Figure 4B, insets). The second phase of the polymyxin E-B5055 ΔlpxM LPS reaction was driven by favorable enthalpic and entropic contributions to binding free energy (Figure 4B, insets).

Figure 4.

Isothermal titration calorimetry (ITC) measurement of the polymyxin E–B5055 LPS binding interactions. (A) K. pneumoniae B5055 LPS. (B) K. pneumoniae B5055ΔlpxM, LPS. Top panels show the heat in µcal/sec per injectant. Bottom panels show the enthalpy (kcal/mol) per injectant at 3 µl of 3 mM polymyxin E into a solution of 0.125 mM LPS in buffer 20 mM HEPES pH 7.0 at 37℃. The insets show the stoichiometry (n), binding affinity (K_d) and thermodynamic parameters (ΔH, enthalpy; TΔS, entropy; ΔG°, free energy) for each interaction derived from the fit of a two independent sites model.²⁰

Assessment of the OM permeabilizing action of polymyxins on K. pneumoniae

NPN uptake by K. pneumoniae

NPN uptake was monitored to assess the permeabilizing activity of polymyxins against the OM of each K. pneumoniae strain. Figure 5A illustrates the increase in fluorescence observed upon the addition of fixed doses of NPN to a suspension of K. pneumoniae cells in the absence of polymyxins. The basal levels of NPN uptake for K. pneumoniae B5055 ΔlpxM strains was relatively higher than the B5055 WT strain, indicative of a more permeable OM structure (Figure 5A). To test the hypothesis that the higher NPN uptake may be a result of the reduced LPS-to-LPS interactions through weaker binding of divalent cations, the effect of a chelating agent, EDTA, on NPN uptake was assessed. The addition of EDTA appeared to increase the uptake of NPN and the maximum level of uptake for the B5055 WT strain (Figure 5A). For the derivative strain B5055 ΔlpxM the effect of EDTA on NPN permeability was far less pronounced, which is consistent with a lower dependency on divalent cation stabilization for their OM integrity (Figure 5A). Figure 5B shows the raw kinetics data for the change in fluorescence observed upon the addition of fixed doses of polymyxin B or E to a suspension of K. pneumoniae cells in the presence of NPN. Polymyxin-induced NPN uptake in the B5055 ΔlpxM strain was more rapid and reached a plateau level faster than occurred for strain B5055. EC₅₀ values documented in Table 1 were obtained from plots of the increase in NPN fluorescence as a function of the polymyxin concentration (Figure 5C). The EC₅₀ values indicated that B5055 ΔlpxM displayed ∼2–3-fold greater sensitivity to polymyxin-induced NPN uptake compared with its parent strain, B5055 (Table 1). The EC₅₀ values showed good correlation with the antibacterial activity of polymyxins estimated from MIC values (Table 1). NPN uptake into suspensions of LPS aggregates was also examined (Table 3). NPN uptake into B5055 LPS aggregate suspensions yielded low peak fluorescence levels (90 ± 3.1; Table 3) compared with the baseline, while higher levels of NPN uptake were observed for LPS aggregates isolated from B5055 ΔlpxM strain (135 ± 8.8; Table 3). Treatments with polymyxins B and E, and EDTA were observed to increase peak NPN fluorescence levels for LPS aggregates from the B5055 WT strain, but not for LPS from B5055 ΔlpxM (Table 3).

Table 3.

Effect of polymyxin B or E on the uptake of NPN into K. pneumoniae LPS aggregates.

K. pneumoniae LPS	Peak NPN fluorescence (arb. units)
K. pneumoniae LPS	NPN per se	EDTA	Polymyxin B	Polymyxin E
B5055	90 ± 3.1	138 ± 4.2	142 ± 6.1	151 ± 4.8
B5055ΔlpxM	135 ± 8.8	140 ± 7.1	139 ± 4.5	144 ± 8.1

Figure 5.

(A) NPN uptake kinetics of a suspension of mid-logarithmic phase K. pneumoniae B5055 WT and B5055ΔlpxM cells in the presence and absence of EDTA. NPN fluorescence emission is shown as a function of time. (B) Effects of polymyxin B on the uptake of NPN into the outer membrane (OM) of K. pneumoniae B5055 WT and B5055ΔlpxM cells. NPN fluorescence emission is shown as a function of time in the presence of increasing concentrations of polymyxin B. (C) The normalized peak NPN fluorescence is plotted as a function of the polymyxin B and E concentration. K. pneumoniae B5055ΔlpxM, polymyxin B (○); K. pneumoniae B5055ΔlpxM, polymyxin E (▪); K. pneumoniae B5055, polymyxin B (□); K. pneumoniae B5055, polymyxin E (◊). Data points are the mean ± SD of three independent measurements.

Lysozyme and deoxycholate permeability assay

Incubation of whole-cell suspensions with deoxycholate, lysozyme or polymyxins alone had a negligible effect on the optical density of the bacterial suspensions (data not shown). However, the addition of polymyxin E to cell suspensions in the presence of lysozyme (Figure 6A) or deoxycholate (Figure 6B) produced a decrease in optical density, which may be interpreted as an increase in deoxycholate- or lysozyme-mediated cell lysis. The lysozyme time course reaction showed that the B5055 ΔlpxM cells were lysed more extensively compared with the B5055 WT cells upon polymyxin E exposure (Figure 6A). Moreover, the B5055 ΔlpxM cells displayed a 20% greater decrease in optical density than the B5055 WT cells at the end of the time course (Figure 6A). Similarly, the deoxycholate permeability assay showed that the B5055 ΔlpxM cells underwent more extensive lysis compared with the B5055 WT cells when exposed to an identical concentration series of polymyxin E (Figure 6B).

Figure 6.

(A) Polymyxin E-induced lysozyme sensitivity. The OD₅₀₀ of a mid-logarithmic phase K. pneumoniae B5055 cell suspensions in the presence of lysozyme following exposure to polymyxin E. (B) Deoxycholate induced cell lysis following exposure to increasing concentrations of polymyxin E. K. pneumoniae B5055 (♦) and B5055ΔlpxM (○). Data are expressed as the percentage of the optical density (OD) in absence of polymyxin E exposure. Data points are the mean ± SD of three independent measurements.

Thermal transition temperature of K. pneumoniae LPS aggregates

Thermal phase transitions in the isolated LPS were monitored using NPN fluorescence as an indicator. Figure 7A shows the phase behavior of LPS from each test strain, the phase-transition temperature and its mid-point (T_m). For the B5055 WT LPS, the gel↔liquid crystalline phase transition of the fatty acyl chains occurred with a T_m of 25℃; in comparison, the B5055ΔlpxM LPS displayed a lower T_m of 23℃ (Figure 7A). In the presence of a fourfold molar excess of polymyxin E to LPS, the T_m of both LPS samples was shifted to lower values (B5055 WT T_m = 23℃; B5055 ΔlpxM T_m = 21℃), indicating increased fluidization of the LPS acyl chains occurred with polymyxin exposure (Figure 7B,C).

Figure 7.

Fluorometric thermal shift measurements of K. pneumoniae B5055 LPS melting profiles. The NPN fluorescence intensity is plotted as function of temperature. (A) Comparison of the K. pneumoniae B5055 LPS melting profiles. (B) Comparison of the thermal shift melting profile of K. pneumoniae B5055 WT LPS (50 µM) in the presence and absence of polymyxin E (200 µM). (C) Comparison of the thermal shift melting profile of K. pneumoniae B5055 ΔlpxM LPS (50 µM) in the presence and absence of polymyxin E (200 µM).

Molecular modeling of the polymyxin B₁–lipid A complexes for each K. pneumoniae strain

To demonstrate the effect of lipid A modifications on the polymyxin binding interaction we constructed molecular models of polymyxin B₁ in complex with the lipid A structure from each K. pneumoniae test strain (Figure 8). The model of the polymyxin B₁–B5055 WT lipid A complexes shows that, in general, the complex is stabilized by a combination of electrostatic and hydrophobic interactions (Figure 8B). The positively-charged side chains of Dab¹ and Dab⁵ bond with the negative charge on the 4′-phosphate head group of lipid A, while those of Dab⁸ and Dab⁹ bond with the 1-phosphate. The buckled configuration of the cyclic peptide portion forces the lipid A binding surface of the polymyxin molecule towards the concave face of the polymyxin molecule. The polymyxin B₁–B5055 ΔlpxM lipid A model suggests that the loss of a secondary 3′-myristate fatty-acyl chain leads to a reduced hydrophobic surface area for interaction with the D-Phe⁶-L-Leu⁷ and the N-terminal octanoic fatty-acyl chain of polymyxin B₁ (Figure 8C). More specifically, the 3′-secondary myristate fatty-acyl chain forms hydrophobic contacts with the octanoic fatty acyl chain of polymyxins. Thus, the effect of deacylation at this position would result in a decreased packing density between the lipid A fatty-acyl chains in the monolayer configuration. Insertion of the polymyxin octanoic fatty-acyl chain into the lipid A leaflet may thus be facilitated in the absence of any steric hindrance from the 3′-myristate, which would account for the increased susceptibility and in vitro binding affinity of polymyxin for LPS from B5055 ΔlpxM compared with B5055 LPS.

Figure 8.

Molecular models of the complex between polymyxin B₁ and the lipid A structure from each K. pneumoniae test strain. The lipid A molecule is shown in space-filling representation and polymyxin B₁ is shown in stick representation. The chemical structure of each represented molecule is shown in the right hand panels. R2, core oligosaccharide with or without attached O-polysaccharide.

Discussion

Increasing reports of resistance to the last-line therapy polymyxins in K. pneumoniae have provided the impetus to elucidate the molecular mechanism(s) of resistance employed by this ‘superbug’. Although Campos et al.²⁹ suggested that susceptibility of K. pneumoniae to cationic antimicrobial peptides was dictated by the thickness of capsular polysaccharide material coating the surface, a recent study revealed that polymyxin susceptibility was not altered in a non-encapsulated K. pneumoniae B5055 mutant, possessing approximately one tenth of the WT capsule levels.¹⁴ Alternatively, structural modifications of lipid A are known to confer polymyxin susceptibility in several other Gram-negative species (reviewed in Groisman³⁰and Peschel³¹); these encompass changes to the lipid A acylation pattern or covalent modification of lipid A phosphates with cationic moieties, such as 4-amino-4-deoxy-L-arabinose or phosphoethanolamine.³² The presence of these cationic substituents in the core or the lipid A structures of heterogeneous LPS preparations can lead to uncertainty in the binding parameters. In the present study we have investigated the influence of the lipid A acylation pattern on the crucial interaction between K. pneumoniae LPS and polymyxins. Nevertheless, under-acylation of the lipid A cannot be attributed as the only reason for the observed differences in the reactivity with polymyxins. We should also remain cognizant of the effects of the heterogeneity and charge properties of the O-antigen and core oligosaccharide components of the K. pneumonia LPS structure (supplementary Table S1). The core polysaccharide consists of a short chain of sugars that is invariably attached to the lipid A component via a Kdo residue through a ketosidic linkage.^33–35 The core is usually conserved across members of a bacterial genus, but can be quite variable between genera of Gram-negative bacteria.^33–35 Thus far, only two core types have been described for K. pneumonia.^36,37 In addition to the common hexoses, the core of the LPS from K. pneumonia serotype O2 contains the Kdo and galacturonic acid (GalA) ‘special’ sugars (supplementary Table S1).^36,37 Unlike the core oligosaccharides of E. coli, Salmonella enterica and Pseudomonas aeruginosa that are rich in negative charges which arise from the phosphate substitution of the Hep sugars;^{34,35,38–42} one noticeable feature of the K. pneumoniae core is the absence of phosphate residues. Instead, the carboxylic acid groups of the Kdo and GalA residues provide the negative charges in their core oligosaccharides.^36–38,43 The O-antigen polysaccharide chain length is highly variable among bacterial strains and can range up to 10–40 repeating units.⁴⁴ So it follows the LPS from a given bacteria strain is a heterogeneous mixture of LPS molecules differing in the length of the O-antigen polysaccharide. The heterogeneity in the number and distribution of O-antigen repeats of LPS samples are seen as a ladder of high molecular mass bands following SDS-PAGE and silver staining (Figure 1). The repeating unit of the K. pneumonia serotype O2 displays a neutral character (supplementary Table S1).

The mutant K. pneumoniae B5055 ΔlpxM strain was derived from strain B5055¹⁴ with a defective lpxM gene which encodes for the late acyltransferase LpxM enzyme responsible for the addition of a secondary lipid A acyl chain. Therefore, compared with the hexa-acylated lipid A predominantly expressed by the B5055 WT, the mutant predominantly expresses a penta-acylated lipid A lacking the 3′-myristate chain.¹⁴ K. pneumoniae isolates defective in LpxM acyltransferase have been shown to be less virulent in animal infection models and produce an LPS with reduced endotoxicity.¹⁴

The bactericidal activity of polymyxin B is known to exhibit a strong temperature dependence.⁴⁵ For example, the activity (i.e. EC₅₀) of polymyxin B against E. coli cells was noted to decrease drastically from 98% at 35℃ to merely 5.4% at 18℃ following treatment with the same polymyxin B concentration.⁴⁵ The temperature region at which the change in killing rate was observed (24–27℃) lies within the phase transition temperature for the E. coli OM, as measured by differential scanning calorimetry.⁴⁵ From a mechanistic view, these observations are reasonable as self-uptake of polymyxins necessitates penetration of the hydrophobic polymyxin domains (i.e. the N-terminal fatty-acyl chain, and D-Phe-L-Leu or D-Leu-L-Leu amino acids of polymyxin B or E respectively) into the lipid A fatty-acyl leaflet.¹³ This integration would certainly be more favorable when the OM is in a fluid liquid-crystalline state. In line with these findings, the results of the fluorometric thermal shift analysis revealed that the B5055 ΔlpxM LPS with a penta-acylated lipid A undergoes a phase transition at lower temperatures compared with the B5055 WT LPS. Accordingly, the penta-acylated lipid A phenotype is likely to provide for an OM leaflet of increased fluidity, as the fatty-acyl layer is less restricted by interchain forces. This observation runs parallel to the increased polymyxin susceptibility of B5055 ΔlpxM in comparison with the B5055 WT strain, as the latter rather promotes a more rigid OM structure owing to stronger lateral interactions between the chains of the hexa-acyalted lipid A species. The importance of the phase state of the OM is readily intelligible based upon the fluorometric thermal shift analysis of mixtures of LPS isolates with fourfold molar excess polymyxin E, which revealed a shift of the fatty-acyl phase transition to lower temperatures. This would suggest that the hydrophobic penetration of polymyxin leads to decreased lipid A interchain interactions, which, in turn, increases the fluidization of the liquid-gel crystalline phase. Collectively, our thermal shift data suggest that the state of order of the lipid A fatty-acyl chains relates to the susceptibility of the K. pneumoniae OM to polymyxins.

Fluorometric and ¹H NMR spectroscopic measurements of the binding affinity of polymyxin B and E to isolated LPS from each strain demonstrated polymyxin B and E bound with greater affinity to the LPS from B5055 ΔlpxM compared with the B5055 WT strain. This would imply that the packing of lipid A fatty acyl chains can considerably affect the polymyxin–LPS binding interaction. This was further highlighted by findings that B5055 ΔlpxM was more susceptible to the lytic action of lysozyme and the anionic detergent, sodium deoxycholate, in comparison with B5055 WT, after exposure to polymyxin E. We propose that these observations may be attributed to a greater propensity for polymyxins to bind and disrupt the OM barrier of B5055 ΔlpxM, thus enhancing the action of these lytic agents.

NPN is a hydrophobic probe whose fluorescence is greatly increased in a hydrophobic environment. Tightly-packed lipid A fatty acyl chains limits the free diffusion of hydrophobic solutes, such as NPN, through the intact OM.¹⁰ However, once permeated, intercalation of NPN into the phospholipid inner leaflet of the OM and the inner cytoplasmic phospholipid membrane bilayer produces a resultant increase in fluorescence. Differences in the kinetics and final NPN uptake levels were observed between B5055 and B5055 ΔlpxM. In comparison with B5055, NPN uptake kinetics were enhanced in B5055ΔlpxM, which corresponds with a less densely packed OM of greater permeability likely arising from the loss of a secondary lipid A acyl-chain in this strain. Further evidence in support of a highly permeable OM comes from the observation that NPN uptake by the B5055ΔlpxM isolate was not affected by the chelating agent, EDTA, which is known to release LPS from the Gram-negative OM.⁴⁶

Conclusions

Our findings provide a molecular-level understanding of the significant role played by the fatty-acyl chains of lipid A for complexation of polymyxin antibiotics by K. pneumoniae LPS. The increased polymyxin susceptibility of B5055 ΔlpxM was paralleled by a greater binding affinity for polymyxins and an increased fluidity of its LPS in comparison with the B5055 WT strain. This may arise from alterations to the LPS-to-LPS fatty-acyl chain packing, which enhances the fluidity of the OM and can account for an increased permeability of B5055 ΔlpxM to NPN. Therefore, the disruption of the K. pneumoniae OM by polymyxins may be partly dependent upon the ability of its N-terminal fatty-acyl chain to penetrate into the fatty-acyl chain layer of lipid A. This event appears to be more facile in strains which predominantly express an under-acylated lipid A with increased OM fluidity. This study provides important insights into the mechanisms of polymyxin activity in this very problematic pathogen.

Footnotes

Funding

R. L. N. and J. L. are supported by research grants from the National Institute of Allergy and Infectious Diseases of the National Institutes of Health (R01A1070896 and R01AI079330). T. V., R. L. N., J. L., P. E. T. and K. R. are also supported by the Australian National Health and Medical Research Council (NHMRC). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Allergy and Infectious Diseases or the National Institutes of Health.

Conflicts of interest

J. L. is an Australian NHMRC Senior Research Fellow. T. V. and M. A. B. are Australian NHMRC Career Development Research Fellows. M. C. is a NHMRC Australia Fellow supported by AF511105.

References

Podschun

Ullmann

. Klebsiella spp. as nosocomial pathogens: epidemiology, taxonomy, typing methods, and pathogenicity factors. Clin Microbiol Rev 1998; 11: 589–603.

Gasink

Edelstein

Lautenbach

Synnestvedt

Fishman

. Risk factors and clinical impact of Klebsiella pneumoniae carbapenemase-producing K. pneumoniae. Infect Control Hosp Epidemiol 2009; 30: 1180–1185.

Walsh

Weeks

Livermore

Toleman

. Dissemination of NDM-1 positive bacteria in the New Delhi environment and its implications for human health: an environmental point prevalence study. Lancet Infect Dis 2011; 11: 355–362.

Sidjabat

Nimmo

Walsh

. Carbapenem resistance in Klebsiella pneumoniae due to the New Delhi Metallo-beta-lactamase. Clin Infect Dis 2011; 52: 481–484.

Kumarasamy

Toleman

Walsh

. Emergence of a new antibiotic resistance mechanism in India, Pakistan, and the UK: a molecular, biological, and epidemiological study. Lancet Infect Dis 2011; 10: 597–602.

Zavascki

Goldani

Nation

. Polymyxin B for the treatment of multidrug-resistant pathogens: a critical review. J Antimicrob Chemother 2007; 60: 1206–1215.

Gales

Jones

Sader

. Global assessment of the antimicrobial activity of polymyxin B against 54 731 clinical isolates of Gram-negative bacilli: report from the SENTRY antimicrobial surveillance programme (2001–2004). Clin Microbiol Infect 2006; 12: 315–321.

Bratu

Tolaney

Karumudi

. Carbapenemase-producing Klebsiella pneumoniae in Brooklyn, NY: molecular epidemiology and in vitro activity of polymyxin B and other agents. J Antimicrob Chemother 2005; 56: 128–132.

Elemam

Rahimian

Mandell

. Infection with panresistant Klebsiella pneumoniae: a report of 2 cases and a brief review of the literature. Clin Infect Dis 2009; 49: 271–274.

10.

Nikaido

. Molecular basis of bacterial outer membrane permeability revisited. Microbiol Mol Biol Rev 2003; 67: 593–656.

11.

Nikaido

. Outer membrane barrier as a mechanism of antimicrobial resistance. Antimicrob Agents Chemother 1989; 33: 1831–1836.

12.

Velkov

Thompson

Nation

. Structure-activity relationships of polymyxin antibiotics. J Med Chem 2010; 53: 1898–1916.

13.

Hancock

REW

Chapple

. Peptide antibiotics. Antimicrob Agents Chemother 1999; 43: 1317–1323.

14.

Clements

Tull

Jenney

. Secondary acylation of Klebsiella pneumoniae lipopolysaccharide contributes to sensitivity to antibacterial peptides. J Biol Chem 2007; 282: 15569–15577.

15.

Wiegand

Hilpert

Hancock

. Agar and broth dilution methods to determine the minimal inhibitory concentration (MIC) of antimicrobial substances. Nat Protoc 2008; 3: 163–175.

16.

EUCAST ECoAST. Breakpoint tables for interpretation of MICs and zone diameters. In: Testing ECoAS. Version 13, 5 January 2011.

17.

Darveau

Hancock

. Procedure for isolation of bacterial lipopolysaccharides from both smooth and rough Pseudomonas aeruginosa and Salmonella typhimurium strains. J Bacteriol 1983; 155: 831–838.

18.

Tsai

Frasch

. A sensitive silver stain for detecting lipopolysaccharides in polyacrylamide gels. Anal Biochem 1982; 119: 115–119.

19.

Lee

Tsai

. Quantification of bacterial lipopolysaccharides by the purpald assay: measuring formaldehyde generated from 2-keto-3-deoxyoctonate and heptose at the inner core by periodate oxidation. Anal Biochem 1999; 267: 161–168.

20.

Soon

Velkov

Chiu

. Design, synthesis and evaluation of a new fluorescent probe for measuring polymyxin-lipopolysaccharide binding interactions. Anal Biochem 2011; 409: 273–283.

21.

Vaara

. Outer membrane permeability barrier disruption by polymyxin in polymyxin-susceptible and -resistant Salmonella typhimurium. Antimicrob Agents Chemother 1981; 19: 578–583.

22.

Helander

Mattila-Sandholm

. Fluorometric assessment of Gram-negative bacterial permeabilization. J Appl Microbiol 2000; 88: 213–219.

23.

Velkov

Lim

Horne

. Characterization of lipophilic drug binding to rat intestinal fatty acid binding protein. Mol Cell Biochem 2009; 326: 87–95.

24.

Pristovsek

Kidric

. Solution structure of polymyxins B and E and effect of binding to lipopolysaccharide: an NMR and molecular modeling study. J Med Chem 1999; 42: 4604–4613.

25.

Pristovsek

Kidric

. The search for molecular determinants of LPS inhibition by proteins and peptides. Curr Top Med Chem 2004; 4: 1185–1201.

26.

Ferguson

Welte

Hofmann

. A conserved structural motif for lipopolysaccharide recognition by procaryotic and eucaryotic proteins. Structure 2000; 8: 585–592.

27.

Bengoechea

Lindner

Seydel

Diaz

Moriyon

. Yersinia pseudotuberculosis and Yersinia pestis are more resistant to bactericidal cationic peptides than Yersinia enterocolitica. Microbiology 1998; 144: 1509–1515.

28.

Brandenburg

David

Howe

. Temperature dependence of the binding of endotoxins to the polycationic peptides polymyxin B and its nonapeptide. Biophys J 2005; 88: 1845–1858.

29.

Campos

Vargas

Regueiro

. Capsule polysaccharide mediates bacterial resistance to antimicrobial peptides. Infect Immun 2004; 72: 7107–7114.

30.

Groisman

. How bacteria resist killing by host-defense peptides. Trends Microbiol 1994; 2: 444–449.

31.

Peschel

. How do bacteria resist human antimicrobial peptides? Trends Microbiol 2002; 10: 179–186.

32.

Raetz

Reynolds

Trent

Bishop

. Lipid A modification systems in Gram-negative bacteria. Annu Rev Biochem 2007; 76: 295–329.

33.

Muller-Loennies

Brade

. Neutralizing and cross-reactive antibodies against enterobacterial lipopolysaccharide. Int J Med Microbiol 2007; 297: 321–340.

34.

Holst

. The structures of core regions from enterobacterial lipopolysaccharides – an update. FEMS Microbiol Lett 2007; 271: 3–11.

35.

Jansson

Lindberg

Wollin

. Structural studies on the hexose region of the core in lipopolysaccharides from Enterobacteriaceae. Eur J Biochem 1981; 115: 571–577.

36.

Vinogradov

Frirdich

MacLean

. Structures of lipopolysaccharides from Klebsiella pneumoniae. Eluicidation of the structure of the linkage region between core and polysaccharide O chain and identification of the residues at the non-reducing termini of the O chains. J Biol Chem 2002; 277: 25070–25081.

37.

Vinogradov

Perry

. Structural analysis of the core region of the lipopolysaccharides from eight serotypes of Klebsiella pneumoniae. Carbohydr Res 2001; 335: 291–296.

38.

Aucken

Pitt

. Different O and K serotype distributions among clinical and environmental strains of Serratia marcescens. J Med Microbiol 1998; 47: 1097–1104.

39.

Knirel Yu

Vinogradov

Kocharova

. The structure of O-specific polysaccharides and serological classification of Pseudomonas aeruginosa (a review). Acta Microbiol Hung 1988; 35: 3–24.

40.

Rocchetta

Burrows

Lam

. Genetics of O-antigen biosynthesis in Pseudomonas aeruginosa. Microbiol Mol Biol Rev 1999; 63: 523–553.

41.

Wilkinson

. Composition and structure of lipopolysaccharides from Pseudomonas aeruginosa. Rev Infect Dis 1983; 5(Suppl. 5): S941–S949.

42.

Knirel

Bystrova

Kocharova

. Conserved and variable structural features in the lipopolysaccharide of Pseudomonas aeruginosa. J Endotoxin Res 2006; 12: 324–336.

43.

Vinogradov

Lindner

Seltmann

. Lipopolysaccharides from Serratia marcescens possess one or two 4-amino-4-deoxy-L-arabinopyranose 1-phosphate residues in the lipid A and D-glycero-D-talo-oct-2-ulopyranosonic acid in the inner core region. Chemistry 2006; 12: 6692–6700.

44.

Stenutz

Weintraub

Widmalm

. The structures of Escherichia coli O-polysaccharide antigens. FEMS Microbiol Rev 2006; 30: 382–403.

45.

Hodate

Bito

. Temperature dependence of bactericidal action of polymyxin B. Microbiol Immunol 1982; 26: 737–740.

46.

Leive

. Release of lipopolysaccharide by EDTA treatment of E. coli. Biochem Biophys Res Commun 1965; 21: 290–296.

Molecular basis for the increased polymyxin susceptibility of Klebsiella pneumoniae strains with under-acylated lipid A

Abstract

Keywords

Introduction

Materials and methods

Materials

Bacterial strains

Determination of minimum inhibitory concentrations

LPS isolation

Kdo purpald assay

Fluorometric measurement of [Dansyl-Lys]1- polymyxin B3 binding to whole cells and isolated LPS

Deoxycholate sensitivity induced by polymyxins

Permeability to lysozyme induced by polymyxins

NPN uptake assay

Fluorometric thermal shift assay

Isothermal titration calorimetry measurement of polymyxin–LPS binding interactions

1H NMR spectroscopy measurements

Molecular modeling of the lipid A–polymyxin B complex

Results

Susceptibility testing and LPS characterization

Binding of polymyxin B and E to isolated K. pneumoniae LPS

Fluorometric assessment of DPmB3 binding to isolated LPS

1H NMR analysis of the binding of polymyxin E with LPS aggregates

Isothermal titration calorimetry assay of polymyxin E binding to isolated LPS

Assessment of the OM permeabilizing action of polymyxins on K. pneumoniae

NPN uptake by K. pneumoniae

Lysozyme and deoxycholate permeability assay

Thermal transition temperature of K. pneumoniae LPS aggregates

Molecular modeling of the polymyxin B1–lipid A complexes for each K. pneumoniae strain

Discussion

Conclusions

Footnotes

Funding

Conflicts of interest

References

Fluorometric measurement of [Dansyl-Lys]¹- polymyxin B₃ binding to whole cells and isolated LPS

¹H NMR spectroscopy measurements

Fluorometric assessment of DPmB₃ binding to isolated LPS

¹H NMR analysis of the binding of polymyxin E with LPS aggregates

Molecular modeling of the polymyxin B₁–lipid A complexes for each K. pneumoniae strain