Physicochemical Characterization of Daptomycin Interaction with Negatively Charged Lipid Membranes.

Daptomycin is known as an effective antibiotic lipopeptide which shows activity against the number of Gram-positive pathogens. Its primary target is the bacterial cell membrane. However, the detailed mechanism of daptomycin action is still subject to debate. In this paper, we have investigated the interactions between lipopeptide and model lipid films composed of negatively charged phosphatidylglycerols and cardiolipin. In order to evaluate the effect of daptomycin on the molecular organization and the properties of lipid assemblies, we have used surface pressure measurements and electrochemical methods combined with atomic force microscopy, quartz crystal microbalance, and surface-enhanced infrared absorption spectroscopy. Our results indicate that daptomycin interaction with the lipid membrane is complex. It involves daptomycin aggregation and partial insertion, which in turn affect the charge distribution on both sides of the membrane and may result in a gradient of water chemical potential. The latter can drive the flux of water across the membrane.

Introduction

Amphiphilic peptides represent a broad class of compounds composed of hydrophobic and hydrophilic amino acid sequences or hydrophilic peptide moieties covalently linked to the lipophilic chains. Because of their amphipathic surfactant-like character, these compounds can organize in numerous supramolecular nanostructures like micelles, fibrils, nanotubes, or nanobelts.[1,2] They also show high affinity to lipid bilayers which makes them potentially useful for biomedical applications. Numerous amphiphilic peptides or lipopeptides were demonstrated to affect the permeability of lipid membranes as a result of pore formation or local solubilization of the membrane.[3−6] For this reason, these compounds are often considered as candidates for new antibiotics, which act on bacterial cell membranes.

Daptomycin is a cyclic lipopeptide produced by the soil bacterium Streptomyces roseosporus. It consists of a cyclic portion composed of 10 amino acid residues and the linear portion including 3 amino acid residues connected to decanoyl fatty acid chain. The characteristic feature of this compound is the presence of nonproteinogenic amino acids: l-kynurenine (Kyn) and l-3-methylglutamic acid (mGlu). It was demonstrated in numerous papers that daptomycin displays a strong antimicrobial activity against the series of Gram-positive pathogens within nanomolar and micromolar range.[7−9] Although the clinical efficacy of daptomycin is quite well established, the molecular mechanism of its action remains unclear. So far it was recognized that daptomycin shows increased affinity to negatively charged membranes containing phosphatidylglycerols, which constitute an important lipid component of the bacterial cell membranes.[10] Moreover, its activity strictly depends on the presence of calcium ions, which can facilitate lipopeptide binding and oligomerization by neutralizing negative charges of peptide moieties and lipid polar heads.[11−13] As a result, daptomycin interacts with the bacterial cell membranes and causes their depolarization.[14−16] The studies on model bilayers revealed increased transmembrane flux of small cations in the presence of daptomycin, which is often interpreted as indicative of pore formation.[17] Nevertheless, some papers have shown results inconsistent with such a scenario. For example, significant perturbation of lipid membrane in giant unilamellar vesicles composed of 1-palmitoyl-2-oleylphosphatidylcholine (POPC) and 1-palmitoyl-2-oleylphosphatidylglycerol (POPG) was observed by Kreutzberger and co-workers.[18] These authors have concluded that daptomycin does not translocate across the lipid bilayer. Instead, it tends to form clusters with negatively charged lipids. An interesting effect of daptomycin on PG-containing bilayers was demonstrated by Chen and co-workers, who observed a lipid-extracting phenomenon, which as the authors claim can be correlated to the antibacterial activity of daptomycin.[19]

To explore the nature of daptomycin action, we have designed model lipid films composed of negatively charged phosphatidylglycerols and cardiolipin, which are known to be the predominant lipid species occurring in cell membranes of Gram-positive bacteria. By using surface pressure measurements and electrochemical methods combined with atomic force microscopy (AFM), quartz crystal microbalance (QCM), and surface-enhanced infrared absorption spectroscopy (SEIRAS), we have examined how the presence of daptomycin affects the electrical and mechanical properties of lipid assemblies. These techniques proved to be useful in molecular-scale characterization of interactions of biologically active compounds with model membranes.[20−23]

Experimental Section

Chemicals

1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphoglycerol (POPG), 1,2-dipalmitoyl-sn-glycero-3-phosphoglycerol (DPPG), and 1′,3′-bis[1,2-dimyristoyl-sn-glycero-3-phospho]-glycerol (CL) were purchased from Avanti Polar Lipids Inc. These particular lipids were chosen as components of lipid films to attain sufficient stability of the monolayers and bilayers. Daptomycin, HEPES, calcium dichloride, and ultrapure methanol were purchased from Sigma-Aldrich, while potassium chloride and chloroform were purchased from POCh Gliwice. Analytical grade sodium hydroxide was purchased from Chempur. All reagents were used as received. The water was purified through Milli-Q system (resistivity 18.2 MΩ × cm). In all experiments, we have used an aqueous buffer solution of HEPES (10 mM), potassium chloride (100 mM), and calcium dichloride (2 mM) adjusted to pH = 7.4 unless otherwise stated.

Surface Pressure Measurements

The monolayers at the air–buffer interface were formed using a KSV NIMA L &LB trough (Biolin Scientific, Sweden) equipped with two movable hydrophilic barriers. HEPES buffer (0.01 M) with 2 mM calcium ions was used as a subphase. Surface pressure was measured using a Wilhelmy plate made of paper. The barriers and the trough were cleaned before each experiment using a chloroform/methanol mixture and then water. Spreading solutions were prepared by dissolving DPPG/POPG/CL (1:1:2 mol/mol/mol) lipids in chloroform/methanol mixture (4:1, v/v) to a final concentration of 1 mg/mL. After spreading, the films at the air–buffer interface were left for 15 min to complete solvent evaporation. Surface pressure versus area per molecule isotherms were recorded at the barriers speed of 10 mm/min and at a constant temperature of subphase (22 ± 1 °C) measured after each experiment. Stock solution of daptomycin was obtained by dissolving lipopeptide in ultrapure water. All experiments at the air–water interface were repeated at least three times to ensure the reproducibility of the results. Calculated errors correspond to the standard deviation.

Liposomes Preparation

Small unilamellar vesicles (SUVs) were prepared according to the procedure described by Barenholtz and co-workers.[24] Stock solutions containing ∼5.0 mg/mL of POPG, ∼5.0 mg/mL of DPPG, and ∼5.0 mg/mL of CL in chloroform/methanol (4:1, v/v) were mixed in a test tube at the desired molar ratio of 1:1:2. The solvent was evaporated by vortexing the solution under the stream of argon, and then the test tube with dry lipid cake was placed in a vacuum desiccator for 1 h. After removal of solvent residues, 1.0 mL of an aqueous buffer solution was added to the lipid cake and the mixture was sonicated at ∼37 °C for 1 h. The resulting suspension of SUVs was transparent.

Electrochemistry

In all electrochemical experiments, we used CHI 750B potentiostat (CH Instruments, Inc., Austin, TX) with a three-electrode cell comprising Ag/AgCl/sat. KCl reference electrode, Pt foil as a counter electrode, and Au(111) as a working electrode. The Au(111) working electrode was cleaned before experiments in piranha solution (H2SO4/H2O2, 3:1 v/v. CAUTION: piranha reacts violently with organic compounds.), rinsed with ultrapure water, and flame-annealed. Self-assembled monolayers of 1-thio-β-d-glucose (further referenced as thioglucose) were prepared by immersing the gold electrode in an aqueous solution containing 0.1 mg/mL of the active compound for ∼2 h, and then the electrodes were thoroughly rinsed with ultrapure water. The electrochemical measurements were carried out in the hanging meniscus configuration. Alternating current (AC) voltammetry measurements were performed with a scan rate of 5 mV/s, AC perturbation with an root-mean-square amplitude of 10 mV, and a frequency of 20 Hz. The differential capacitance curves were derived from the in-phase and out-of-phase components of the AC signal under the assumption that the electrode–electrolyte interface can be considered as a simple RC circuit. In order to determine potential of zero free charge, the immersion method was used.[25] Briefly, the working electrode was placed above the electrolyte and polarized to the desired potential (the applied potentials were in the range of +0.30 to −0.40 V with 0.05 V interval). Then the electrode was slowly lowered toward the electrolyte and the current flow was measured upon contact. The same method was utilized in chronocoulometry measurements. Electrochemical impedance spectroscopy experiments were performed within the frequency range of 10–1 to 104 Hz at the potential of +0.1 V and the amplitude of 0.01 V. All of the measurements were performed at 22 ± 1 °C. All potentials reported in this work are referenced to Ag/AgCl/sat. KCl electrode. The supporting electrolyte used in all electrochemical experiments was an aqueous buffer solution of HEPES (10 mM), potassium chloride (100 mM), and calcium dichloride (2 mM) adjusted to pH = 7.4.

Quartz Crystal Microbalance with Dissipation Monitoring

Q-Sense E1 instrument (Q-Sense AB, Sweden) was used for real-time monitoring of mass and viscoelasticity changes of the lipid membranes immobilized on solid support upon the addition of daptomycin. The sensor crystals were 5 MHz, AT-cut quartz electrodes with 100 nm evaporated gold and titanium adhesion layer (QSX 338). Before the experiments, the crystals were cleaned with piranha solution (concentrated H2SO4/H2O2) (3:1 v/v) for approximately 5 min, rinsed thoroughly with Milli-Q water, then sonicated in 2% Hellmanex solution for 30 min in 35 °C and rinsed with Milli-Q water. After drying under Ar stream, the crystals were immersed in 0.1 mM thioglucose solution in methanol for 12 h. After that, the substrates were rinsed with methanol and water and mounted in an electrochemical three-electrode QCM-D cell with miniaturized Ag/AgCl/sat. KCl as a reference electrode and platinum wire as a counter electrode. Potential was controlled with CHI 750B potentiostat (CH Instruments, Inc., Austin, TX). HEPES buffer with 2 mM calcium ions was passed through the chamber using a peristaltic pump at the flow rate of 0.15 mL/min for about 15 min to obtain the stable baseline of frequency and dissipation. Solid-supported lipid bilayers were prepared by a liposome spreading method. The stock solution of DPPG/POPG/CL liposomes was injected into the QCM-D chamber at a flow rate of 0.15 mL/min. Following the formation of a stable lipid bilayer, crystals were rinsed with buffer solution to remove any nonruptured vesicles. After approximately 20 min, the 3 μM solution of daptomycin was added at the flow rate of 0.15 mL/min. Afterward, the pump was stopped. QCM-D crystals covered with lipid bilayer were subjected to a stagnant daptomycin solution up to 3 h. The values of frequency and dissipation were measured at four overtones (third, fifth, seventh, and ninth).

Topography Imaging

Atomic force microscopy (AFM) experiments were performed with Dimension Icon (Bruker) in PeakForce Tapping Mode with ScanAsyst Fluid probes (Bruker, nominal spring constant 0.7 N/m). The samples were imaged under in situ conditions in an aqueous buffer solution at the temperature of 22 ± 1 °C. Potential was controlled with CHI 750B potentiostat (CH Instruments, Inc., Austin, TX), platinum wire served as a counter electrode and miniaturized Ag|AgCl|sat.KCl as a reference electrode The bilayers were deposited on thioglucose-modified Au(111) substrates by spreading small unilamellar vesicles and the bilayer formation was completed within approximately 2 h. The thickness of lipid bilayers was determined based on a cross-sectional analysis of defects as an average height difference between the bare substrate and the region covered by lipid film.

Surface-Enhanced Infrared Absorption Spectroscopy

The spectra were recorded using Nicolet iS50 infrared spectrometer (Thermo Fisher Scientific Inc.) equipped with liquid nitrogen cooled MCT-A detector and custom-made single-reflection accessory. The incident angle was set at 60°, and the spectral resolution was 4 cm–1. The spectra are displayed in absorbance units defined as A = log(I0/I), where I0 and I correspond to the intensities of IR radiation observed for the reference and the sample, respectively. In all experiments, we have used all-glass spectroelectrochemical cell with platinum foil as a counter electrode and Ag/AgCl/sat. KCl as a reference electrode. The working electrode was a thin gold film deposited on a reflectance plane of a silicon hemispherical prism accordingly to the procedure described in the literature.[26] PGSTAT101 potentiostat (Metrohm Autolab) was utilized to control electrode potential. Omnic 9 software (Thermo Fisher Scientific Inc.) was used for data processing.

Results and Discussion

Surface Pressure Measurements

To assess the aggregation behavior of daptomycin under the experimental conditions used in this work, we have determined its critical micelle concentration (CMC). Figure presents the dependence of the surface pressure as a function of the natural logarithm of daptomycin concentration in an aqueous buffer solution of HEPES (10 mM), potassium chloride (100 mM), and calcium dichloride (2 mM) adjusted to pH = 7.4. At low concentrations, the surface pressure changes sharply with increasing concentration of lipopeptide. This is related to increasing number of surfactant molecules at the interface. However, above a certain concentration the dependence becomes weak since daptomycin interfacial concentration does not change significantly. Consequently, we can observe two linear regimes on the plot and the inflection point indicates CMC, which was found to be 91 μM. For comparison, the value of 0.12 mM was determined from fluorescence measurements in 0.1 M KCl by Kirkham and co-workers.[27] The small difference may result from the different composition of the solution as well as the method used for CMC determination since surface tension measurements often exhibit lower sensitivity compared with fluorescence-based assays. To ensure complete dispersion of molecules in solution, the concentration of daptomycin in subsequent experiments was kept at the level of well below CMC that is within the range of 0.5–3 μM.

Figure 1

Dependence of the surface pressure as a function of the natural logarithm of the bulk concentration of daptomycin. Surface pressure measurements were performed on an aqueous buffer solution of HEPES (10 mM), potassium chloride (100 mM), and calcium dichloride (2 mM) adjusted to pH = 7.4.

According to numerous reports, bacterial cell membrane is considered as a primary target for daptomycin, which is assumed to incorporate into the lipid assembly and affect the overall permeability of the bilayer.[28] In order to evaluate daptomycin interactions with lipids, we have performed surface pressure measurements utilizing POPG/DPPG/CL monolayer spread at the air–buffer interface. The ability of daptomycin to incorporate into the lipid film was examined by following the surface pressure changes of the monolayer during its lateral compression on an aqueous HEPES buffer containing calcium ions and dissolved lipopeptide (see Figure ). The lift-off of the isotherm recorded for POPG/DPPG/CL monolayer in the absence of daptomycin starts at area per molecule equals approximately 150 Å2. Then the phase transition occurs that is attributed to the transition from a liquid-expanded (LE) to a liquid condensed phase (LC). After the phase transition, the isotherm rises rapidly while reducing the area per molecule and collapses at approximately 62 mN/m. This confirms the formation of a stable and well-ordered monolayer at the air–buffer interface. Next, the POPG/DPPG/CL isotherms were recorded on the buffered subphase containing three different concentrations of daptomycin: 5 × 10–7 M, 10–6 M, and 3 × 10–6 M. The analysis of isotherms allows one to determine characteristic parameters of the monolayers, that is, limiting and minimum molecular area as well as collapse surface pressure. The limiting molecular area was determined by extrapolating the slope of the isotherm to zero surface pressure, and it is interpreted as the hypothetical area occupied by one molecule in either solid or liquid-condensed phase. The minimum molecular area refers to the area at which collapse pressure occurs, that is, the smallest area occupied by a molecule in the monolayer.

Figure 2

Surface pressure versus area per molecule isotherms (A) and the changes in compression modulus as a function of the surface pressure (B) obtained for POPG/DPPG/CL monolayers compressed on pure buffer subphase (black curve) and buffer subphase containing different concentration of daptomycin (colored curves). The aqueous buffer solution contained HEPES (10 mM), potassium chloride (100 mM), and calcium dichloride (2 mM) and it was adjusted to pH = 7.4.

The obtained isotherms depicted in Figure demonstrate that the presence of daptomycin in the subphase induces changes in the surface properties of POPG/DPPG/CL membrane. The lift-off point of the isotherms shifts toward higher values of area per molecules with increasing daptomycin concentration. Similar changes are observed for limiting area per molecule and minimal area per molecule (Table ). This suggests that daptomycin incorporates into POPG/DPPG/CL membrane during its formation and it is not squeezed out from the monolayer even at high surface pressure values. However, the increasing amount of daptomycin within the POPG/DPPG/CL membrane decreases the stability of the monolayer which is manifested by the decreasing value of the surface pressure at which the collapse of the monolayer occurs. Moreover, for the lowest concentration of daptomycin in the subphase the phase transition shifts toward a slightly higher value of surface pressure and it becomes less developed, while it completely disappears for higher daptomycin concentration. Such behavior together with increasing limiting area per molecule is commonly attributed to the increasing fluidity of the model lipid membrane.[29] These changes on isotherms are also manifested in a variation of the values of compression modulus, which is defined aswhere π is the surface pressure and A is the area per molecule. This parameter provides information on the state in which the monolayer exists at a given surface pressure. It is assumed that compression modulus below 12.5 mN/m corresponds to a gaseous phase, between 12.5 and 100 mN/m, the monolayer is in a liquid expanded state, the range of 100–250 mN/m is characteristic for a liquid condensed state, and the values above 250 mN/m are indicative of a solid state. The maximum value of compression modulus for the monolayer registered on pure buffer subphase is equal to 163 mN/m, which indicates that the monolayer is in a liquid-condensed state. The addition of daptomycin to the subphase leads to a slight decrease in the maximum of compression modulus indicating some disordering effect on lipid film. This effect becomes more pronounced with the increasing concentration of lipopeptide. The maximum value of compression modulus, although decreased, remains in the range corresponding to the liquid-condensed phase. Hence, the daptomycin substantially decreases molecular packing density within the film but with moderate fluidizing effect. This may be indicative of relatively strong lipid–lipopeptide interactions, which hinder to some extent motional freedom of the molecules within the assembly. On the other hand, substantially increased molecular area might be a relevant factor contributing to the increased permeability of the film.

Table 1

Properties of Monolayers at the Air–Buffer Interface

monolayer	limiting area per molecule [Å2]	collapse pressure [mN/m]	minimum area per molecule [Å2]	maximum Cs–1 [mN/m]
POPG/DPPG/CL (1:1:2)	72.0 ± 0.1	62.8 ± 0.8	32.4 ± 0.9	163 ± 9
+ 5 × 10–7 M daptomycin	116.5 ± 0.4	56.8 ± 0.9	78.9 ± 0.8	165 ± 12
+ 10–6 M daptomycin	169.4 ± 1.2	49.2 ± 0.9	124.5 ± 0.7	151 ± 10
+ 3 × 10–6 M daptomycin	277.4 ± 0.9	39.7 ± 0.9	215.6 ± 1.1	140 ± 7

The above-presented results concern the action of daptomycin during the formation of a model lipid membrane. However, in physiological conditions the antibiotic interacts with already existing cell membranes. Therefore, we have examined the effect of daptomycin on the already precompressed lipid layer (Figure ). The POPG/DPPG/CL monolayer was first compressed to the surface pressure of 35 mN/m, since at this value the molecular packing within the monolayer roughly corresponds to the packing of the molecules in natural cell membranes.[29] The surface pressure of a pure POPG/DPPG/CL monolayer measured at a constant area decreases slightly over time. This decrease may result from the partial solubility of the lipids in the subphase or from the oxidation of the double bond present in POPG molecule.[30] Moreover, since all three components of the monolayer are negatively charged and the monolayer at 35 mN/m exists in a highly ordered state, the repulsive forces between molecules may induce the expulsion of some molecules into the subphase. After compression of POPG/DPPG/CL monolayer to 35 mN/m, daptomycin was injected into the subphase to give the desired final concentration and the changes in surface pressure were recorded in time until equilibrium pressure was reached. The injection of daptomycin results in a substantial rapid increase in surface pressure, which is mainly driven by interactions between daptomycin–Ca2+ complex and negatively charged lipid polar headgoups and the incorporation of the drug into the lipid monolayer due to the hydrophobic interactions. However, after reaching the maximum a slight decrease in surface pressure was observed, which may result from a slow reorganization of lipid chains upon interactions with drug molecules.[31] The quasi-plateau indicating that the monolayer approaches the steady-state was reached after approximately 45–60 min. The surface pressure increases steadily with the increasing daptomycin concentration in the subphase. The analysis performed 3 h after the addition of daptomycin revealed the increase in surface pressure values by 10.6, 14.0, and 19.8 mN/m for daptomycin concentration equals 5 × 10–7, 10–6, and 3 × 10–6 M, respectively. Such a significant and stable increase in surface pressure indicates that daptomycin interacts with lipid polar headgroups and inserts into the hydrophobic region of the lipid layer. These results clearly indicate that daptomycin can spontaneously insert into negatively charged POPG/DPPG/CL membrane.

Figure 3

Changes in a surface pressure as a function of time recorded for POPG/DPPG/CL monolayer compressed to initial surface pressure of 35 mN/m on an aqueous buffer after injection of daptomycin into the subphase. An aqueous buffer solution contained HEPES (10 mM), potassium chloride (100 mM) and calcium dichloride (2 mM) and it was adjusted to pH = 7.4.

Electrochemical Characterization

Although the monolayers at air–water interface are often used as a model assembly to evaluate the interactions between lipids and biologically relevant molecules, they do not fully reflect supramolecular organization of lipids within biological membranes. Therefore, we have evaluated the daptomycin behavior in the presence of model lipid bilayer supported on solid surface. For that purpose, POPG/DPPG/CL bilayer was deposited on gold electrode premodified with thioglucose forming a floating bilayer lipid membrane (fBLM).[32,33] Such approach offers several advantages. First of all, it enables electrochemical control of the transmembrane potential by applying external electric field of comparable magnitude as in biological membranes. Another advantage of fBLM is that thioglucose interlayer can act as hydrophilic cushion which assures better hydration of lipid polar heads located in the inner leaflet of the membrane in the vicinity of electrode surface. Recently, the presence of water layer between POPG/DPPG/CL membrane and thioglucose-modified gold electrode was proved experimentally using surface-enhanced infrared absorption spectroscopy.[34]

To evaluate how the stability of the lipid membrane is affected by the electric field, we have performed differential capacitance measurements. As can be concluded from Figure , the variation of the differential capacitance for intact fBLM is minor between +0.4 and −0.3 V with a shallow minimum at ∼4.7 μF/cm2, indicating that the membrane is stable within this potential range. The capacitance is growing at more negative potentials, which reflect the increasing amount of water accumulating at the interface between the bilayer and the thioglucose. The latter is related to electroporation of the film and the onset of electrodewetting.[35] Finally, the pseudocapacitive peak at −0.73 V appears, which is indicative of the desorption of the membrane.

Figure 4

Potential-dependent changes in a differential capacitance of thioglucose-modified gold electrodes with POPG/DPPG/CL membrane. The supporting electrolyte was an aqueous buffer solution of HEPES (10 mM), potassium chloride (100 mM), and calcium dichloride (2 mM) adjusted to pH = 7.4. Inset shows schematic view of the fBLM architecture.

Further, we have used the immersion method to determine the potential of zero free charge (Epzfc) for thioglucose-modified gold electrodes with POPG/DPPG/CL membrane in the absence and in the presence of daptomycin. Figure illustrates the resulting plots of the charge density as a function of the potential applied to the electrode. It can be seen that the value of Epzfc for intact POPG/DPPG/CL membrane is about +0.18 V, while in the presence of daptomycin it is shifted down to −0.20 V.

Figure 5

Free charge plotted as a function of the immersion potential for thioglucose-modified gold (111) electrode with the POPG/DPPG/CL floating bilayer in the absence (black) and presence (red) of daptomycin (3 μM). Each point represents an average from three independent measurements and error bars correspond to the standard deviation. The supporting electrolyte was an aqueous buffer solution of HEPES (10 mM), potassium chloride (100 mM), and calcium dichloride (2 mM) adjusted to pH = 7.4.

Daptomycin is known to bind with calcium ions with 1:1 stoichiometry.[13] It means that the resulting complex is still monoanion, which after binding to the membrane can alter the charge distribution across the membrane and contributes to the negative shift in Epzfc. It is supported by the results of chronocoulometry experiment shown in Figure performed within similar time-scale as liposome leakage experiments reported in the literature. In this case, the potential of +0.10 V was applied to the electrode modified with intact fBLM and then the electrode was brought into the contact with supporting electrolyte. Since Epzfc for this system was determined to be +0.18 V, the potential of +0.10 V is negative with respect to Epzfc. Therefore, the electrode surface becomes negatively charged under these conditions as can also be deduced from negative charge flow. Using the value of Epzfc, one can roughly estimate the transmembrane potential as a difference between potential applied to the electrode and the potential of zero free charge (ϕtrans = Eapp – Epzfc).[35] The resulting value of ϕtrans is −0.08 V, which is close to the transmembrane potentials observed in biological membranes. Further, the daptomycin was injected into the electrochemical cell and approximately after 5 min of exposure, the Q versus t curve displayed well-pronounced inflection and the charge started to change toward the zero value. The latter is reached after approximately 20 min upon injection of lipopeptide. Such behavior may be considered as indicative of the slow dissipation of transmembrane potential resulting from the redistribution of ions on both sides of the membrane. This seems to be in line with earlier reported observations that daptomycin causes incomplete or at least slow depolarization of the membrane.[36] After prolonged exposure to daptomycin, the charge was still rising and finally reached positive values. It means that the electrode surface became positively charged. This is fully in line with the results shown in Figure , since in the presence of the daptomycin the charge density at the metal surface is positive at the potential of +0.1 V. These results demonstrate that indeed daptomycin has a strong effect on the distribution of ions on both sides of the lipid membrane, which in turn affects the charge of the electrode surface.

Figure 6

Charge versus time curves recorded after immersion of the thioglucose-modified Au(111) electrode with POPG/DPPG/CL floating bilayer in a buffer solution while holding constant potential of +0.1 V. Black curve was recorded in the absence of daptomycin, while red curve was recorded under the same conditions with the exception that daptomycin was added to the electrochemical cell ∼12 min after immersion of the electrode. The concentration of lipopeptide was 3 μM. The supporting electrolyte was an aqueous buffer solution of HEPES (10 mM), potassium chloride (100 mM), and calcium dichloride (2 mM) adjusted to pH = 7.4.

In order to evaluate whether the effect of membrane depolarization is related to pore-forming activity of daptomycin, we have performed electrochemical impedance spectroscopy (EIS) experiments. EIS is a reliable method for detection of lipid membrane-damaging effects triggered by pore-forming proteins or peptides. It was demonstrated by Valincius and co-workers that formation of the pores causes the decrease in the measured impedance and the phase angle displays well-pronounced minimum in the low frequency region.[37] Moreover, the logarithm of the frequency corresponding to the said minimum can be correlated with the density of the pores. The results of the EIS measurements performed with POPG/DPPG/CL fBLM are shown as Bode plot in Figure . We have observed that the impedance decreases upon injection of the daptomycin indicating increased permeability of the membrane. However, the phase angle plot does not display a minimum in low frequency region. Hence, this process does not seem to be accompanied by formation of the stable pores within the lipid membrane. Moreover, the decrease of impedance is temporal and after 15 min of exposure its value increases above the initial level. This is commonly interpreted as sealing of the membrane. However, in this case it might be related to decreased water content within the interfacial region between the electrode and the bilayer.

Figure 7

Bode plots presenting logarithm of impedance (squares) and phase angle (circles) as a function of the frequency for POPG/DPPG/CL fBLM on gold electrode before (black curve) and after (colored curves) addition of 3 μM daptomycin. Spectra were collected at the potential of +0.1 V. The supporting electrolyte was an aqueous buffer solution of HEPES (10 mM), potassium chloride (100 mM), and calcium dichloride (2 mM) adjusted to pH = 7.4.

Quartz Crystal Microbalance Studies

For further evaluation of daptomycin action on negatively charged fBLM, we have used quartz crystal microbalance which enabled monitoring of the changes in a frequency (Δf) and energy dissipation (ΔD) of a gold-coated quartz crystal sensor covered by thioglucose interlayer and lipid membrane. These two parameters allow a conclusion about the variation in mass and viscoelastic properties of the solid-supported lipid film.[38,39] The variation in Δf is indicative of either mass loss or gain depending whether the changes are positive or negative, respectively. The variations in energy dissipation are related to the changes in the rigidity of the membrane immobilized on the sensor. Namely, an increase of ΔD reflects the softening of the film, while its stiffening is characterized by a decrease of ΔD values. Moreover, by analyzing the frequency and dissipation changes at different overtones one can conclude about the spatial distribution of mass and rigidity changes on the sensor surface.[40,41] The response of the sensor at higher overtones is sensitive to those processes, which occur close to the sensor surface, while the response at the lower overtones brings information about the changes at the film-water interface. Figure illustrates how Δf and ΔD changes in time upon exposure of POPG/DPPG/CL bilayer to daptomycin at 3 μM concentration at the potential of +0.10 V (i.e., ϕtrans = −0.08 V). We can observe substantial frequency drop accompanied by an increase in dissipation energy. Such response of the sensor is indicative of daptomycin aggregation, which is accompanied by the decreased coupling of the lipid membrane with the sensor. The latter could be explained by daptomycin-induced interruption of the ordering of lipids resulting from decreased molecular packing density. Interestingly, the QCM-D data show clearly that the variations in Δf and ΔD at lower overtones are definitely larger compared to higher overtones. It means that the changes occurring within the lipid film are more pronounced in the upper leaflet of the bilayer compared to the region in the vicinity of the sensor surface.

Figure 8

Quartz crystal microbalance frequency Δf (solid lines) and energy dissipation ΔD (dashed lines) response recorded for thioglucose-modified gold with POPG/DPPG/CL floating bilayer in the presence of 3 μM daptomycin (A) and corresponding (ΔD) vs (Δf) plot (B). Lipopeptide was injected into the cell approximately after 23 min of stabilization of the system (indicated by red arrow in panel A). The supporting electrolyte was an aqueous buffer solution of HEPES (10 mM), potassium chloride (100 mM), and calcium dichloride (2 mM) adjusted to pH = 7.4.

More detailed information on the mechanism of daptomycin action can be obtained from ΔD versus Δf plots shown in Figure B. Such presentation of the QCM-D can be used as a fingerprint for the lipopeptide-membrane interaction. Initially the linear growth of ΔD accompanied by the decrease of Δf values is observed. This is indicative of mass accumulation with corresponding loss of the membrane rigidity and it is interpreted as an adsorption of lipopeptide on the membrane. This process is very rapid but it slows down after ∼30–40 s. Then, the maximum is observed at ΔD versus Δf plots followed by small decrease of ΔD. This is indicative of the change in the mechanism of daptomycin-membrane interaction and it can be interpreted as slow insertion of the lipopeptide into the lipid bilayer. Clearly, the upper leaflet is more affected by this process since the changes at third overtone are definitely much more pronounced compared with higher overtones. On the basis of this observation, we may assume that lipopeptide aggregates predominantly in the upper part of the membrane and then some fraction inserts into the lipid film but its spatial distribution across the bilayer is nonhomogenous. Another important conclusion from QCM-D experiments is that the daptomycin does not cause rupture or dissipation of the membrane. As shown in Figure A, the quasi steady state is reached after ∼20–30 min, and then the mass as well as viscoelastic properties of the assembly does not vary considerably indicating that there are no further changes in a membrane structure/organization.

Atomic Force Microscopy Imaging

Daptomycin aggregation was also verified using atomic force microscopy (AFM). This technique enables monitoring the changes in the topography of POPG/DPPG/CL bilayers supported on thioglucose-modified Au(111) electrode after exposure to lipopeptide. The imaging was performed under in situ conditions. First, the samples were imaged in the absence of lipopeptide to ensure that the bilayers are stable during repetitive scans. Figure A presents an exemplary image of intact fBLM at the potential of +0.10 V (i.e., ϕtrans = −0.08 V). Lipid membrane is homogeneous and covers almost all scanned area. Small membrane spanning defect is visible in the top section of the image. There are also some topographically lower regions visible at the borders between the domains where the molecules are less ordered. The thickness of the lipid membrane was estimated as an average height difference measured across the defect sites. Cross-sectional analysis of the images obtained for five independent samples revealed that the mean thickness of the lipid film is 5.2 ± 0.5 nm. This value is in good agreement with previous studies reporting AFM imaging of lipid bilayers.[42] Then, an aqueous stock solution of daptomycin was injected to reach the final concentration of lipopeptide of 3 μM and the imaging was continued.

Figure 9

(A) AFM images presenting intact POPG/DPPG/CL bilayer deposited on thioglucose-modified Au(111) electrode; (B) POPG/DPPG/CL bilayer immediately after injection of daptomycin; (C) POPG/DPPG/CL bilayer 60 min of exposure to daptomycin. All images were acquired in an aqueous buffer at the potential of +0.10 V. The concentration of daptomycin was 3 μM. Roman numerals indicate different phases detected after exposure of the membrane to daptomycin. The supporting electrolyte was an aqueous buffer solution of HEPES (10 mM), potassium chloride (100 mM), and calcium dichloride (2 mM) adjusted to pH = 7.4. Cross-sectional profiles taken across the images are shown in Figure S2 in Supporting Information.

It was found that the exposure of POPG/DPPG/CL bilayer to daptomycin results in the formation of the elongated aggregates which appear on top of the membrane immediately after lipopeptide injection (see Figure B). Dimension analysis revealed that the length of aggregates varies from 50 nm up to 300 nm, while the width is between 15 and 40 nm. The average height of the aggregates is 6.1 ± 0.6 nm. This dimension corresponds roughly to the doubled length of daptomycin molecule, which is about 3.2 nm. Therefore, we propose that aggregates can be described as the assemblies where lipopeptide molecules are organized into the double-layered structure with partially interdigitated lipophilic chains to compensate for the large size of the polar heads. Such organization of lipopeptides was reported in several research papers.[43−45] At this point, it should also be stressed that the concentration of daptomycin used in all experiments is well below critical micelle concentration and the molecules are fully dispersed in a buffer solution. Therefore, we conclude that assembly of daptomycin on lipid membrane is a surface-induced phenomenon driven by lipid–lipopeptide interactions. This statement is supported by the results of reference experiments, that is, the imaging of thioglucose-modified gold under the same experimental conditions did not reveal the formation of aggregates. Moreover, we have observed that aggregation phenomena on POPG/DPPG/CL bilayer do not occur in the absence of Ca2+. This proves that aggregation depends on the synergistic effect of negatively charged lipids and calcium ions. We believe that our AFM results agree very well with the findings reported by Lee and co-workers.[46] On the basis of circular dichroism (CD) studies, these authors have demonstrated that daptomycin is in the monomeric form before membrane binding at therapeutic micromolar concentrations and it has a unique CD state independent of calcium ion concentrations. However, the daptomycin has another unique CD state after binding to PG-containing membranes with calcium ions, possibly in a form of oligomers or aggregates bound with lipids. Nevertheless, the partitioning of some fraction of daptomycin molecules into the fBLM cannot be excluded, as can be deduced from QCM-D results. Continuous AFM imaging of the POPG/DPPG/CL membrane revealed that the structure of the assemblies evolves in time and noticeable fraction of daptomycin aggregates merge with lipid bilayer. The resulting film morphology is shown in high-resolution image in Figure C, where three different phases can be distinguished. Phase I is ascribed to the area where the bilayer morphology seems to be unchanged. Phase II indicates the domain, which is on average ∼0.4 nm lower compared to phase I and it displays visible undulations. Finally, phase III has similar morphology as phase II but this region is ∼1.0 nm lower compared to phase I. Such changes in morphology of the lipid membrane indicate that certain population of lipopeptide molecules is released from aggregates and inserts into the lipid bilayer. As can be deduced from AFM image shown in Figure C daptomycin insertion affects the ordering of the molecules within lipid assembly since the thickness of the film is decreased. The observed corrugation may indicate that after insertion of daptomycin, the molecules within the film try to rearrange to accommodate lipophilic chains of lipopeptide and compensate for its bulky head groups. As a result, certain curvature is induced locally in the membrane giving rise to undulations. This observation seems to be in line with the postulate that daptomycin induces positive curvature strain on the lipids.[14] Assuming that such a scenario is valid, phase II can be ascribed to the bilayer domain where daptomycin insertion affected predominantly the upper leaflet, while phase III corresponds to a more disordered region where some fraction of the lipopeptide is present in both leaflets. Such interpretation would be in line with QCM-D fingerprints, which demonstrated that lipopeptide inserts into the membrane to some extent, however, noticeably larger fraction seems to be accumulated in the upper part of the bilayer. Another important information gained from AFM imaging is that the continuity of the membrane is preserved even upon daptomycin insertion. Hence, the AFM results support the conclusion that membrane disruption does not take place after daptomycin aggregation. This is in line with the findings reported by Cotroneo and co-workers who demonstrated that daptomycin bactericidal activity against Staphylococcus aureus is accompanied by negligible cell lysis and no membrane discontinuities have been detected by electron microscopy.[47]

In order to get molecular-scale insight into the mechanisms of daptomycin action on negatively charged fBLM, we have employed surface-enhanced infrared absorption spectroscopy (SEIRAS).[48,49] This method takes advantage of the fact that infrared absorption intensity can be significantly enhanced by 10–1000 times on nanoparticles of coinage metals due to the polarization of metal nanoparticles by the electromagnetic field of the incident light. The latter induces an oscillating dipole in the metal nanoparticle through excitation of localized plasmon. Further, the induced dipole produces an electric field around the nanoparticle and excites the absorbed molecules. The induced electric field is normal to the local metal surface, which implies that only those molecular vibrations where transition dipole moment has component normal to the local surface can be excited and enhanced. On the basis of such surface selection rule, one can conclude about the orientation of adsorbed molecules with respect to the surface. Importantly, the local electric field decays within a short distance away from the surface. It means that local enhancement occurs only in the vicinity of the metallic islands and a dominant contribution comes from the species close to the metal surface. For this reason, SEIRAS can provide information about the amount and the structure of interfacial water. Figure illustrates the differential spectra collected for POPG/DPPG/CL fBLM during its exposure to daptomycin action, while the potential applied to the gold electrode was held at +0.1 V. The spectra recorded before injection of the lipopeptide was used as reference. Daptomycin binding to the membrane is confirmed by relatively well-developed bands located at ∼1653 and ∼1531 cm–1. The position of these bands correlates very well with amide I and amide II bands observed in daptomycin IR spectra. Interaction of daptomycin with lipid membrane resulted in emergence of negative ν(OH) band with global minimum at ∼3350 cm–1 indicative of loss of hydrogen-bonded water.[50] This band is most likely accompanied by a weak negative δ(HOH) band near 1645 cm–1, but in this case it is probably overlapped with relatively strong positive amide I band from daptomycin. Nevertheless, the appearance of the negative ν(OH) band demonstrates that lipopeptide aggregation leads to the partial dehydration of the lipid polar heads. Additionally, the loss of water from the region between the membrane and thioglucose may also contribute to the emergence of the negative ν(OH) band, since the SEIRAS is more sensitive to the changes in the direct vicinity of the electrode. Such interpretation is in line with the results of EIS measurements where the decrease of the capacitance of the submembrane region was observed. Hence, the negative ν(OH) band could be related to the efflux of water molecules from the interfacial region between the membrane and thioglucose interlayer to the outer plane of the bilayer and/or bulk of the solution. Daptomycin initially adsorbs on top of the membrane and lipopeptide molecules may displace water bound to phosphates and carbonyl groups. Hence the chemical potential of water in the polar head region of the upper leaflet is lower compared to the polar head region of the proximal leaflet. This in turn may result in a situation similar to that when the membrane experiences osmotic stress.[51] As a consequence, the gradient of chemical potential drives the flux of water molecules across the membrane. It should also be noted that daptomycin binding resulted in the emergence of positive bands related to ν(C–H) and ester ν(C=O) stretching vibrations. Both may originate from aggregated daptomycin, or alternatively, indicate the change in the orientation of lipids. In the latter case, it would imply that lipid molecules are tilting upon lipopeptide binding and the membrane becomes thinner or more disordered, which is consistent with QCM-D and AFM results. Such behavior is reasonable in the view of the previously discussed effect of water replacement by lipopeptide. The latter may affect the orientation of the lipids, which in turn locally changes their packing and can lead to the formation of defects facilitating the passage of water.

Figure 10

Time evolution of SEIRA spectra recorded at the potential of +0.1 V for thioglucose-modified gold with POPG/DPPG/CL floating bilayer upon exposure to 3 μM daptomycin. The reference spectrum was recorded for the same system before injection of daptomycin. The supporting electrolyte was an aqueous buffer solution of HEPES (10 mM), potassium chloride (100 mM), and calcium dichloride (2 mM) adjusted to pH = 7.4.

Conclusions

In this paper, we have used a multitechnique experimental approach to follow how daptomycin interacts with a negatively charged membrane composed of POPG/DPPG/CL (1:1:2). Surface pressure measurements with monolayers at the air–buffer interface demonstrated that daptomycin shows the ability to bind and insert into the lipid films. Simultaneously, it changes the packing density of the molecules causing moderate fluidization of the membrane manifested by an increase in mean molecular area and decrease of compression modulus. This observation was confirmed by further experiments under electrochemical control for fBLMs immobilized on gold electrodes premodified with thioglucose. As revealed by QCM-D data obtained at ϕtrans = −0.08 V, daptomycin aggregates quickly on top of the lipid membrane and this process is followed by slow insertion of the lipopeptide into fBLM. This leads to the softening of the film, however, the overall change in mass and the viscoelastic properties is more pronounced for an upper leaflet, which demonstrates that larger fraction of the lipopeptide accumulates in the upper part of the bilayer. AFM imaging fully confirms such a scenario, since the formation of aggregates on top of the fBLM is observed immediately after daptomycin injection. Further changes in bilayer morphology strongly suggest possible partitioning of the lipopeptide into the membrane. Interestingly, electrochemical measurements revealed that the exposure of fBLM to daptomycin at ϕtrans = −0.08 V leads to a substantial redistribution of the charge resulting in membrane depolarization. It is accompanied by the directional flux of water from the submembrane region to the bulk as proved by SEIRAS measurements. On the basis of the analysis of our results, we have concluded that lipopeptide mode of action is complex and initially it involves relatively fast adsorption of daptomycin in a form of aggregated assemblies. At this stage, lipopeptide binds to the lipids and concomitantly displaces water molecules bound to the polar heads. This process has multiple consequences. First of all, the charge distribution across the membrane is changed since the excessive negative charge is accumulated in the upper leaflet the fBLM. This way the potential gradient across the membrane is lowered. Second, the binding of daptomycin affects the orientation of lipids and molecular ordering. The latter facilitates lipopeptide insertion, and as a consequence the membrane curvature pronounced as corrugation is produced. Such local changes in packing can lead to the formation of defects which increase the permeability of the membrane for water and small ionic species. Moreover, since the water molecules are displaced during lipopeptide binding in the upper leaflet, it can be expected that a certain gradient of water chemical potential across the membrane is created,[45] which in turn drives the temporal efflux of water from the submembrane region. Nevertheless, the temporarily increased permeation of water/ions is not associated with the formation of stable pores but rather stems from local disordering of the lipid film. Such a scenario seems to contradict the generally accepted model of pore-forming activity of daptomycin. However, it should be noted that our model system contains a relatively large fraction of palmitoyl lipids, which do not support membrane permeabilization.[52] Also cardiolipin, which is present in our model system, was demonstrated to prevent daptomycin translocation into the inner leaflet of the membrane.[53] This may explain the lack of pore-forming activity and preferential accumulation of daptomycin in the outer leaflet of the membrane observed in this work. Nevertheless, our results show that even in the presence of palmitoyl lipids and cardiolipins daptomycin is still able to depolarize the membrane and temporarily increase its permeability. Moreover, the observation of daptomycin-induced changes in the orientation of lipids and/or molecular ordering may have important implications for natural membranes since the localization of proteins involved in cell division and cell wall synthesis may be affected this way.