Atomic Force Microscopy and Electrochemical

0 downloads 0 Views 3MB Size Report
Oct 17, 2014 - It is often considered as a good model for general studies on interactions of antimicrobial .... bilayers were immersed in 10 mM or 1 mM solution of melittin. At ... substrates were placed in the electrochemical cell and the lipid membrane ... This suggests that the film formed on gold electrode is not perfectly ...

Electrochimica Acta 162 (2015) 53–61

Contents lists available at ScienceDirect

Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta

Atomic Force Microscopy and Electrochemical Studies of Melittin Action on Lipid Bilayers Supported on Gold Electrodes Joanna Juhaniewicz 1, Slawomir Sek * ,1 Faculty of Chemistry, Biological and Chemical Research Centre, University of Warsaw, Zwirki i Wigury 101, 02-089 Warsaw, Poland

A R T I C L E I N F O

A B S T R A C T

Article history: Received 2 July 2014 Received in revised form 9 October 2014 Accepted 10 October 2014 Available online 17 October 2014

Melittin is an amphipathic helical peptide which shows strong membranolytic activity against bacterial cells as well as human erythrocytes. It is often considered as a good model for general studies on interactions of antimicrobial peptides with biological membranes since the detailed mechanisms of their action on biological membranes are still subject of intense debate. In this paper we have used electrochemical methods combined with in situ AFM imaging in order to evaluate the mechanisms involved in melittin membranolytic activity. We have observed that high concentration of the peptide causes rapid degradation of a single component DMPC bilayer supported on gold electrode. The lipid membrane undergoes micellization and subsequent dissolution. This indicates that under such conditions melittin acts according to detergent-like mechanism. The mode of melittin action differs substantially when the peptide concentration is lowered. In this case, the changes in DMPC bilayer structure are less rapid and at the initial stages the peptide adsorbs on top of the membrane. This process is followed by fluidization of the DMPC film, which facilitates further reorientation and insertion of melittin into the bilayer. As a result, the permeability of the membrane is increased. AFM data shows that sharp differentiation between carpet and toroidal pore mechanism is difficult and it is very likely that melittin acts according to mixed mechanism. Yet different behavior of melittin was observed for the mixed DMPC/Cholesterol bilayer. The susceptibility of the membrane to melittin action was reduced probably due to the increased packing within the hydrocarbon chain region of the bilayer. We have also noticed that the fluidization of the membrane is a common feature for all systems studied here. Thus, it seems to be a crucial step for melittin action which enables the peptide molecules to adopt proper orientation either for pore formation or disruption of the membrane. ã 2014 Elsevier Ltd. All rights reserved.

Keywords: Supported lipid membranes Atomic force microscopy Gold electrodes Bioelectrochemistry Antimicrobial peptides

1. INTRODUCTION Rapidly increasing number of multi-drug resistance pathogens resulted in a continuous need for new active compounds with strong activity. One of the most promising groups of compounds includes antimicrobial peptides [1]. These are small naturally occurring peptides with molecular mass less than 10 kDa and usually positively charged. Antimicrobial peptides are functional substances in the innate immune system of virtually all forms of life and they constitute a first line of defense against pathogens [2]. They display a broad range of activity on bacteria, fungi and parasites. Therefore, they have enormous potential as novel therapeutic agents. Based on the targets of AMPs action, they can be divided in two major groups: membrane disruptive AMPs

* Corresponding author. Tel.: +48 225526661. E-mail address: [email protected] (S. Sek). ISE Member

1

http://dx.doi.org/10.1016/j.electacta.2014.10.039 0013-4686/ ã 2014 Elsevier Ltd. All rights reserved.

which cause the increase in membrane permeability and nonmembrane disruptive AMPs which can pass through the membrane to reach the intracellular targets and inhibit the synthesis of cell wall, proteins, RNA or DNA [3]. Irrespective of the targets of AMPs action, the key step is always their interactions with cell membrane driven by electrostatic or hydrophobic forces. Until now, there are four main models describing the possible way of interactions of AMPs with membranes [4–6]. In the barrel-stave model peptides bind to the membrane surface as monomer, next reorient perpendicularly to the surface during the insertion process and, after reaching the threshold peptide–lipid (P/L) ratio, form barrel-stave-like aggregates with the hydrophobic regions interacting with hydrophobic tails of membrane and hydrophilic regions facing each other and forming the transmembrane pore or channel. In this case, lipids are not involved in a pore structure and the membrane does not bend during formation of the pore. The latter is the biggest difference between barrel-stave and toroidal model, in which hydrophilic regions of peptide remain in contact with polar lipid head groups during the pore formation. As a

54

J. Juhaniewicz, S. Sek / Electrochimica Acta 162 (2015) 53–61

consequence, membrane bends sharply inward to form a pore lined by both lipids head groups and peptides. The third, micellaraggregate model assumes that after insertion into lipid membrane, peptides associate into micelle-like complexes without any fixed stoichiometry. Such aggregates are formed in a concentration- and voltage-dependent manner and contribute to the disruption in permeability. This way transmembrane transport of the molecules including lipids and peptides is enabled [7,8]. Different model, that is carpet model, involves adsorption of the peptides parallel to the membrane surface and covering the membrane in a carpet-like manner. When the threshold peptide concentration on the surface is reached, the membrane breaks down and lipid aggregates surrounded by peptide molecules are formed. The membrane disruption, therefore, is not associated with peptide insertion and pore formation; it is rather dissolved by peptide in a dispersionlike fashion. Melittin is an example of membrane active peptide. It is watersoluble amphipathic helical peptide consisting of 26 amino acids, which is isolated from the honeybee Apis mellifera. The overall net charge of melittin molecule is +6 and most of the charge is localized at the C-terminus of the peptide [9]. Melittin interacts with biological as well as artificial membranes. For example, it causes hemolysis of cells and leakage of entrapped dyes in lipid vesicles. Melittin is a non-selective peptide. In other words, it is a toxin that displays strong lytic activity against bacterial cells as well as human erythrocytes [10]. This feature precludes melittin for a direct therapeutic use but it offers a great model for general studies on interactions of antimicrobial peptides with biological membranes. The mode of action of melittin depends on the properties of lipid bilayer, therefore numerous aspects have to be considered, including structural requirements of peptide, the length and unsaturation of lipid acyl chains, charge of polar headgroup, role of cholesterol, melittin concentration and its spatial orientation in lipid membranes [11,12]. Supported lipid membranes constitute an attractive model of biological membrane. Therefore, such systems are often used as biomimetic architectures, which enable investigation of various membrane-related biological processes [13–15]. Properties of solid supported lipid membranes are determined by number of factors including the lipid composition, size, shape and presence of unsaturated bonds. Equally important is the nature of the substrate. Most widely used are glass, quartz, silicon and mica [16–18]. However, metal surfaces were demonstrated to be useful as well [19]. Importantly, they show certain advantage, since the structure of the supported film as well as lipid-lipid and lipidprotein interactions can be probed in full electrochemical conditions. This enables investigation of voltage-dependent membrane processes including structural changes as well as lipid-lipid and lipid-protein interactions. In this work, we demonstrate an electrochemical approach combined with in situ AFM imaging, which enabled the evaluation of the mode of melittin action in the presence of artificial lipid membranes supported on gold electrodes. Electrochemical methods were demonstrated to be useful for monitoring the structural changes in supported lipid bilayers occurring upon action of membrane disrupting peptides [20–23]. We use cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) in order to verify the permeability of intact lipid films and then after their exposure to melittin. AFM imaging allowed in situ observation of structural changes within the bilayers. It is well known that the modification of the electrode with blocking dielectric layer leads to significant decrease of the capacitance accompanied by increase of charge transfer resistance [24]. The latter results from the fact, that the electrode surface cannot be directly accessed by redox probe. Consequently, electron transfer between electroactive species and the electrode surface is inhibited and it occurs

from certain distance defined by the thickness of the immobilized film. As it was already mentioned, action of melittin affects the structure of lipid membranes. This should result in significant changes in a permeability of the membrane to ions and water molecules. Since capacitance as well as charge transfer resistance are sensitive to the presence of the defects in a bilayer, both could be used for evaluation of the mechanism of melittin action. Using such approach, we have investigated how the mode of melittin action is determined by the nature of the lipid membrane. For this purpose, single-component membrane composed of 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC) was compared with two-component system containing DMPC and cholesterol. DMPC is broadly used for preparation of biomimetic membranes. It was chosen as a model system since the structure of gold supported bilayers of this particular lipid are very well characterized using photon polarization modulation infrared reflection absorption spectroscopy (PMIRRAS) [25,26], atomic force microscopy (AFM) [27], and neutron reflectivity (NR) [28]. Cholesterol is a common component of the membranes of eukaryotic cells [29]. It affects the properties of the lipid membranes, providing enhanced stiffness and rigidity due to the increased packing density within the hydrophobic region [30]. Since the natural target for melittin is the membrane of erythrocyte that contains up to 30 mol% of cholesterol [10], the understanding of interactions of melittin with membrane cholesterol becomes an important issue. Several studies have revealed that the role of cholesterol in melittin activity depends on the lipid composition: in case of homogeneous lipid bilayers (i.e. composed of a single lipid component), cholesterol inhibits the action of melittin [31,32], whereas in heterogeneous systems like lipid rafts or a mixture of zwitterionic lipids cholesterol does not provide any additional protection against melittin [33,34]. 2. EXPERIMENTAL All chemicals were purchased from Sigma-Aldrich with exception of 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC), which was purchased from Avanti Polar Lipids Inc. The distilled water used in all experiments was passed through a MilliQ water purification system and its final resistivity was 18.2 MV/ cm. Monolayers were prepared at the air-water interface using a KSV LB trough 5000 (KSV Ltd., Finland). Before each experiment, trough and barriers were washed using the mixture of chloroform and methanol and finally rinsed with Milli-Q water. The compression of the monolayers was performed at the barriers speed of 10 mm/min at a constant temperature of 22  1 C. The spreading solutions of lipids were prepared by dissolving DMPC and cholesterol samples in chloroform. Lipid bilayers were transferred on gold substrates (11 11 mm slides, Arrandee), which were 200-300 nm thick gold films evaporated onto borosilicate glass precoated with a 4 nm thick adhesive layer of chromium. Prior to the deposition of the lipid bilayer, gold substrates were cleaned in the mixture of H2O2/NH3/H2O (1:1:5 v/ v/v) at 70 C, rinsed with water, dried and then flame annealed. The latter involved heating of the gold electrode in a propane/butane flame until the dark red glowing was observed. Then the substrate was quenched with water and the procedure was repeated for 67 times. As a result, large Au(111) terraces were obtained with the length of the edge in the order of few hundred nanometers. For atomic force microscopy experiments gold beads prepared according to Clavilier procedure were used [35]. The quality of both types of gold substrates was evaluated using AFM and the representative images are shown in Figure 3S in supporting information. Lipid bilayers were transferred from the air-water interface (with pure water as a subphase) onto the solid supports at the pressure of 35 mN/m. This particular pressure was chosen

J. Juhaniewicz, S. Sek / Electrochimica Acta 162 (2015) 53–61

because the lateral pressure of the natural membranes has been estimated to be around 35 mN/m [36,37]. The first lipid layer was deposited by vertical withdrawal of the substrate at the speed of 25 mm/min for DMPC monolayers and 5 mm/min for DMPC/CHOL (7:3) monolayers. The substrates were left to dry for approximately 1.5 hour and the second lipid layer was transferred at the surface pressure of 35 mN/m using the horizontal touch technique. Electrochemical measurements were performed using CHI 750B bipotentiostat (CH Instruments Inc., Austin, TX) in a threeelectrode cell with Ag/AgCl/sat.KCl as a reference electrode and platinum foil as a counter electrode. Potassium hexacyanoferrate (III) and potassium hexacyanoferrate (II) were used as electroactive probes. The electrochemical impedance spectra were fitted to Randles equivalent circuit using EIS Spectrum Analyzer [38]. The action of melittin on lipid membranes was studied in a buffer solution (20 mM TRIS, 150 mM NaCl and 5 mM EDTA, pH = 7.5  0.1), E = 0.26 V, DU = 5 mV. Gold substrates modified with lipid bilayers were immersed in 10 mM or 1 mM solution of melittin. At 10 mM concentration, melittin exists as a monomer in aqueous solution [39]. In the presence of 150 mM NaCl, melittin was shown to exhibit a strong response on DMPC membrane [40]. In order to diminish melittin binding to DMPC bilayer and to slow down its self-association, we have also studied the interactions between melitin and DMPC bilayer upon decreasing the concentration of the peptide down to 1 mM. It was reported that melittin concentrations of 0.5 to 2.0 mM induced release of membrane permeability markers without complete disruption of the cell membrane [41]. Melittin concentrations of 10 mM and 1 mM represent two different regimes of the peptide action. The incubation time was always 15 minutes. After this time, the modified substrates were carefully rinsed with water and transferred to the electrochemical cell. Between each measurement, the electrodes were stored in a pure buffer. Atomic force microscopy (AFM) images were obtained with Dimension Icon (Bruker). The temperature was carefully controlled during the experiments and it was kept at 22  1  C. The imaging was performed in a buffer solution (20 mM TRIS, 150 mM NaCl and 5 mM EDTA, pH = 7.5  0.1). Prior to each experiment, the electrochemical cell was cleaned in piranha solution (concentrated H2SO4/30% H2O2 3:1, v/v) for at least 2 hours and rinsed thoroughly with pure water. The images were taken in soft tapping mode using PPP-BSI cantilevers (Nanosensors) with the nominal spring constants between 0.01 N/m and 0.1 N/m. Since the PPP-BSI tips are designed for contact mode, we have carried out control experiments in order to verify their performance in tapping mode operation. For the details see supporting information. The amplitude of free oscillation of the cantilever was usually in the range of 15-20 nm, while the amplitude setpoint was between 10 and 13 nm. The AFM images were collected at the scan rates of 24 Hz with 512  512 points resolution. The bilayer-modified substrates were placed in the electrochemical cell and the lipid membrane was imaged several times to prove that no disruption of the membrane occurs due to the imaging. After that, the samples were incubated in melittin solutions following the same procedure as for electrochemical measurements. All images were taken under the same conditions as reported for electrochemical experiments, however, in the absence of the redox probe. The image processing was carried out using Pico Image Basic 5.1.1 software (Agilent Technologies) and involved first order line by line leveling. In some cases spline filtering was used. The thickness of the lipid films was determined from force-distance curves using the procedure described in the literature [42]. For that purpose we have used point and shoot procedure in which particular spot on the surface is first selected on AFM image and then force spectroscopy measurement is performed. Exemplary force-distance curves as well as distributions of the determined thicknesses for all systems

55

are provided as supporting information. The thickness measured from each individual force-distance curve was corrected for elastic deformation based on Hertzian model [43]. The errors reported for the thicknesses are standard deviations. 3. RESULTS AND DISCUSSION 3.1. DMPC Bilayer vs. 10 mM Melittin The influence of 10 mM solution of melittin on DMPC bilayer organization was examined in the presence of ferricyanide/ ferrocyanide redox couple. Fig. 1A presents exemplary cyclic voltammograms obtained for DMPC bilayer supported on gold electrode before (0 min) and after exposure to melittin solution (15 and 30 minutes). In the absence of melittin the peak-to-peak separation is approximately 550 mV for the redox reaction of ferricyanides, instead of the 60 mV observed on bare gold electrode. It proves that electron transfer process at the modified electrode is slowed down in the presence of the bilayer and the direct access to the electrode surface is hindered. With increasing time of melittin action, the separation of peaks decreases significantly and is equal to 320 mV and 200 mV for 15 and 30 minutes, respectively. This confirms destabilizing effect of melittin on DMPC bilayer. The latter becomes less compact and the electrode surface becomes more accessible for ions diffusing from the bulk of the solution. This way, the electron transfer is facilitated, which is reflected by decreasing peak-to-peak separation. Enhanced electron transfer efficiency may also result from electrostatic interactions. In other words, melittin adsorbed at the

Fig. 1. Cyclic voltammetry (A) and electrochemical impedance spectra for DMPC bilayer immobilized on gold electrode, before (black curve) and after exposure (colored curves) to 10 mM melittin solution recorded in the presence of 5 mM Fe (CN)63 /4 . Time of exposure is given in insets.

56

J. Juhaniewicz, S. Sek / Electrochimica Acta 162 (2015) 53–61

Table 1 Electrochemical characteristics of DMPC bilayer exposed to 10 mM melittin. Time of exposure to melittin

Cdl (mF/cm2)

RCT (kohm/cm2)

DEp (mV)

0 min 15 min 30 min

1.84  0.11 3.97  0.18 9.72  0.42

4.47  0.27 2.31  0.12 0.49  0.03

550 320 200

bilayer surface attracts electrostatically negatively charged redox probe and therefore may also provide its easier access to the gold electrode. Similar conclusions can be drawn based on impedance

potential range from -0.4 to + 0.3 V vs. SSCE. The images presented in this work were acquired at the potential of +0.26 V vs. Ag/AgCl, therefore the existence of ripples is reasonable. In our case, the periodicity of the ripples was 6.0  1.0 nm and the amplitude was 1.0  0.5 nm. There are also defect sites present within the film. Their depth distribution is bimodal with average values of 2.7  0.3 nm and 4.3  0.5 nm (see Figure 5S in supporting information). It can be concluded that there are certain populations of the defects which span single leaflet or entire bilayer. The presence of the latter explains slightly increased value of the measured capacitance. The thickness of the intact DMPC film

Fig. 2. AFM images (400  400 nm2) of DMPC bilayer transferred onto the Au(111) electrode by combined Langmuir-Blodgett and Langmuir-Schaefer techniques. (A) Image of intact DMPC bilayer; (B) image of the same bilayer after 15 minutes exposure to 10 mM melittin; (C) image of the same bilayer after 30 minutes exposure to 10 mM melittin. Images were taken at the same spot on the sample and were recorded in a TRIS buffer (pH=7.4) at the potential of +0.26 V vs. Ag/AgCl.

spectra, since the charge transfer resistance becomes smaller as observed by the reduction of the semicircle diameter with increasing time of melittin action (see Fig. 1B). After 30 minutes of melittin action, charge transfer resistance drops by one order of magnitude (see Table 1). The decrease of the charge transfer resistance is accompanied by the increase of the capacitance. Similar trend was observed by Brevnov and Finklea [44]. Initial value of capacitance is slightly higher than typical capacitances for biological membranes that range between 0.8 and 1.0 mF cm 2 [45]. This suggests that the film formed on gold electrode is not perfectly uniform but has some defects or pinholes filled with electrolyte solution, which contribute to the increase of the mean dielectric constant of the system. However, upon 30 minutes of exposure to melittin, we have observed up to five-fold increase of capacitance. This may indicate that melittin molecules either affect the dielectric properties of the bilayer system or cause membrane thinning. Relatively high final value of the capacitance (10 mF/ cm2) brings us to conclusion that the continuity of the lipid bilayer is disrupted and the bare gold is exposed to the electrolyte solution locally. Thus, the electrochemical data strongly suggest that melittin solubilizes large fragments of the bilayer in a detergent-like manner. In order to verify conclusions based on electrochemical results, we have performed in situ AFM imaging of DMPC bilayer supported on gold in the absence as well as upon addition of melittin. Fig. 2A demonstrates the AFM image of intact DMPC bilayer on gold electrode. The most striking feature is the presence of the undulation, which resembles ripple phase (Pb’). As it was demonstrated by Li and coworkers, appearance of such structure results from the fact that the polar headgroup of DMPC is larger than cross section of two hydrocarbon chains and to compensate for the size mismatch, lipid molecules are packed with their polar groups alternatively displaced in a sawtooth manner [27]. These authors observed that the undulation was stable within the

Fig. 3. Cyclic voltammetry (A) and electrochemical impedance spectra for DMPC bilayer immobilized on gold electrode, before (black curve) and after exposure (colored curves) to 1 mM melittin solution recorded in the presence of 5 mMFe (CN)63 /4 . Time of exposure is given in insets.

J. Juhaniewicz, S. Sek / Electrochimica Acta 162 (2015) 53–61

57

3.2. DMPC Bilayer vs. 1 mM Melittin

Table 2 DMPC bilayer subjected to 1 mM melittin. Time of exposure to MLT

Cdl (mF/cm2)

RCT (kohm cm2)

DEp (mV)

0 min 30 min 60 min 240 min 12 h

1.92  0.12 1.19  0.09 1.57  0.10 3.76  0.16 3.99  0.11

4.21  0.19 5.33  0.21 4.32  0.17 2.19  0.14 1.98  0.10

570 * 650 380 330

determined from force-distance curves was found to be 4.5  0.6 nm. Fig. 2B presents the same bilayer after 15 minutes exposure to 10 mM melittin solution. As can be seen, the electrode surface is covered by a large number of spherical aggregates with average diameter about 20 nm. Thus, during 15 minutes of exposure, the melittin was able to fluidize the membrane, penetrate it and form mixed micelles. However, at this stage the micelles are still attached to the gold substrate. Further imaging revealed that aggregates are removed gradually from the surface and large patches of bare gold are exposed to the solution (dark areas). This explains very well our electrochemical results. The electrode surface upon partial removal of the lipid film becomes accessible for redox probe and solvent molecules, contributing to lower charge transfer resistance and rise of the capacitance. Direct visualization using AFM indicates clearly that at 10 mM concentration the peptide solubilizes large fragments of the bilayer in a detergent-like manner. Similar conclusions have been drawn by other authors, who studied melittin interactions with phosphatidylcholine lipids and observed the dissolution of the membrane in the presence of similar or much higher concentrations of melittin covering the range from 10 mM up to 200 mM [20,46].

It is known that at high concentration many membrane-active peptides display detergent-like behavior, although their mechanism of action is different at lower concentration [47]. In order to verify whether the melittin also behaves this way, we have performed the measurements at decreased melittin concentration equal to 1 mM. Indeed the changes in the membrane permeability in the presence of 1 mM melittin are substantially different from those observed at 10 mM melittin. Fig. 3A shows the cyclic voltammograms recorded in the presence of 5 mM Fe(CN)63 /4 using gold electrode modified with DMPC bilayer. The curve obtained for intact bilayer displays large separation between anodic and cathodic peaks which is approximately 550 mV. Thus, again we observe that the electrode reaction in the presence of the bilayer is inhibited due to the limited access of the redox probe to the electrode surface. Surprisingly, after 30 minutes of the interaction with melittin, the currents were even lower and the peaks corresponding to reduction/oxidation of electroactive probe are strongly suppressed. Such behavior demonstrates that the electrode reaction is further inhibited by the presence of melittin. However, after 60 minutes, peaks can be observed again and they become well pronounced and less separated with increasing time of melittin action. Similar changes were observed in impedance spectra. During the first hour after the immersion of DMPCmodified electrode in melittin solution, the charge transfer resistance increases and the capacitance decreases compared with the pure lipid membrane. This indicates that the permeability of the bilayer is reduced (Fig. 3B). After two hours, the charge transfer resistance started to decrease and finally, after four hours it was about two times lower comparing with intact lipid

Fig. 4. AFM images (250  250 nm2) of DMPC bilayer transferred onto the Au(111) electrode by combined Langmuir-Blodgett and Langmuir-Schaefer techniques. Image (A) shows intact DMPC bilayer, while images (B); (C) and (D) show the same sample after 30, 120 and 240 minutes exposure to 10 mM melittin respectively. The average thickness of the film is shown in a bottom right corner of each image. The cross marks in panel (B) correspond to the points which are representative for the values of the thickness shown in the AFM image. All images were taken at the same spot on the sample and were recorded in a TRIS buffer (pH=7.4) at the potential of +0.26 V vs. Ag/AgCl.

58

J. Juhaniewicz, S. Sek / Electrochimica Acta 162 (2015) 53–61

membrane. Then, it remained practically constant for next several hours (Table 2). Similarly, initial decrease in the double layer capacitance upon exposure of DMPC bilayer to melittin was followed by further increase of its value up to 4 mF/cm2 after several hours. Such behavior can be explained by initial adsorption of melittin molecules on the surface of the bilayer and further reorientation. At low concentration, melittin first adsorbs on the membrane and simultaneously reduces its permeability for ions and water. It results in increased charge transfer resistance and decreased capacitance. Similar behavior of supported lipid membranes in the presence of 1.6 mM melittin was described by Steinem and coworkers [20]. However, these authors reported that after 10 minutes the capacitance as well as the conductivity of the membrane reached constant value. In our experiment, we were able to observe decrease of the charge transfer resistance with a simultaneous increase of the capacitance approximately within 60-90 minutes after the adsorption. Such changes reflect the situation where the melittin molecules start to reorient from parallel to perpendicular orientation and incorporate into the lipid membrane. This penetration results in the formation of the defects in the DMPC bilayer that facilitate the transport of the ions and water molecules to the electrode surface. The system seems to achieve the equilibrium after approximately 120 min because further interactions of melittin with DMPC bilayer have practically no effect on both resistance and capacitance values. This indicates that 1 mM melittin disturbs the membrane, probably by inducing formation of pores and/or larger defects which only locally disrupt

Fig. 5. Cyclic voltammetry (A) and electrochemical impedance spectra for DMPC/ Cholesterol (7:3) bilayer immobilized on gold electrode, before (black curve) and after exposure (colored curves) to 10 mM melittin solution recorded in the presence of 5 mMFe(CN)63 /4 . Time of exposure is given in insets.

membrane continuity. After 12 hours of melittin action, the capacitance of DMPC bilayer is 4 mF/cm2, which is still much lower than value observed for bare gold electrode (24 mF/cm2). It strongly suggests that the electrode surface is still covered by the dielectric layer established by lipid film but surely it contains numerous defects. To evaluate whether the scenario given above is reasonable, we have performed in situ AFM imaging of DMPC bilayer in the presence of 1 mM melittin. Initial structure of the membrane was exactly the same as previously, with long range order as indicated by the presence of the ripples (see Fig. 4A). The initial thickness of intact DMPC bilayer was 4.4  0.4 nm. In the presence of melittin the morphology of the film starts to change. After 30 minutes of melittin action we can observe the accumulation of aggregates on the top of the bilayer. These are represented by bright patches in Fig. 4B and most likely correspond to peptide molecules adsorbed on top of the bilayer. The distribution of the film thickness is bimodal in this case (see Figure 9S in supporting information). The average thickness measured at the region represented by the red mark in Fig. 4B was 6.9  0.4 nm, while for the white mark it was 4.4  0.4 nm. In other words the aggregates are higher by around 2.5 nm, which is in a good agreement with topographical data extracted from AFM image (i.e. about 2-3 nm height difference). Since the preferred orientation for the peptide molecules is parallel to the membrane plane, we can conclude that the aggregates are composed of more than one layer of melittin. Due to the adsorption of the peptide on top of the bilayer the transport of electroactive ions across the membrane is hindered. Therefore it would explain the suppression of the electron transfer and the decrease of the capacitance observed in electrochemical experiments. As parallel mode of action, direct adsorption of the melittin molecules at the defect sites can also be considered. This would contribute to the increased charge transfer resistance as well. Interestingly, adsorption of melittin is accompanied by a disappearance of ripples. This is attributed to the fact that peptide may induce the phase transition within the lipid bilayer and the latter becomes more fluid. Importantly, the fluidization may facilitate further reorientation and insertion of the peptide molecules into the lipid membrane. It is confirmed by the image taken after 120 minutes of melittin action shown in Fig. 4C. In this case, the aggregates of peptide disappeared indicating that melittin is already inserted into the membrane. At this stage, the film is strongly disordered. Its thickness decreased significantly comparing to Fig. 4B. Now the value determined from force-distance curves is 3.8  0.4 nm, which is also noticeably lower than initial thickness of DMPC membrane. The thinning of the film suggests that indeed it is more fluid comparing with intact DMPC bilayer. Longer exposure to melittin results in partial recovery of the ripples as demonstrated in Fig. 4D. Nevertheless, the undulation is less pronounced and quite irregular comparing with initial structure of DMPC membrane. Long-term imaging did not reveal any further changes in morphology of DMPC film, thus the image shown in Fig. 4D represents the equilibrium state observed in electrochemical experiments. The final thickness of the DMPC film is 4.3  0.4 nm. The sequence of the AFM images confirms that at lower concentration the melittin does not induce micellization of the membrane. It rather involves initial aggregation of the peptide on top of the DMPC bilayer followed by the fluidization of the membrane. The latter facilitates reorientation and insertion of the melittin into the lipid film, which most likely leads to formation of pores and also larger defects. Nevertheless, the membrane remains attached to the gold substrate as opposite to previous case where the removal of the film was observed. 3.3. DMPC/Cholesterol Bilayer vs. 10 mM Melittin Electrochemical experiments performed with DMPC/Cholesterol bilayer were similar to those presented for DMPC

J. Juhaniewicz, S. Sek / Electrochimica Acta 162 (2015) 53–61 Table 3 DMPC/Cholesterol bilayer subjected to 10 mM melittin. Time of exposure to MLT

Cdl (mF/cm2)

RCT (kohm cm2)

DEp (mV)

0 min 30 min 60 min 120 min 180 min

0.65  0.02 1.04  0.07 1.46  0.08 1.92  0.15 3.84  0.27

15.0  0.92 9.18  0.34 5.78  0.23 4.21  0.28 1.88  0.08

** ** ** 470 380

membranes. First, the permeability of DMPC/Cholesterol bilayer was verified using cyclic voltammetry in the absence of melittin. As can be seen in Fig. 5A, the addition of cholesterol dramatically improved the blocking properties of the membrane since we cannot observe any peaks. In other words, diffusive currents were completely suppressed and only small exponential rise of the current with increasing overpotential was observed. Interestingly, after 60 minutes of melittin action, the transport of the redox probe through the membrane is still strongly inhibited as indicated by the absence of the peaks. However, the current is slightly higher comparing with unimpaired DMPC/Cholesterol bilayer. Thus, the electron transfer is more efficient but the membrane is still impermeable for mass transport. Slightly higher currents observed here may result from the partial penetration of the bilayer by the redox probe. It should be noted that the presence of 30% cholesterol increases the space between DMPC headgroups [48]. Therefore, melittin may percolate directly into the polar region of the bilayer without an intermediate step involving adsorption on top of the membrane. This enables also closer approach of the negatively charged redox probe, since it could be electrostatically

59

attracted by cationic melittin which is already partially embedded within the membrane. As a consequence, electron transfer between electroactive probe and electrode surface occurs from shorter distance giving rise to higher currents. Longer exposure of DMPC/Cholesterol bilayer to melittin action leads to the membrane disordering, which enables direct mass transport to the gold electrode. However, the peak-to-peak separation observed after 180 minutes is approximately 350 mV, which demonstrates that melittin disturbs the structure of membrane only locally, most likely in a pore forming manner. The evaluation of the membrane permeability based on impedance measurements brings us to the same conclusions. The presence of the DMPC/Cholesterol bilayer immobilized on gold electrode impedes the diffusion of redox probe to the gold electrode resulting in high charge transfer resistance. As demonstrated in Table 3, also substantial decrease in double layer capacitance can be observed, which is indicative of the presence of the compact layer with a relatively low dielectric constant. Moreover, the values calculated for DMPC/Cholesterol bilayer are different from those estimated for a single component DMPC bilayer by the factor of 3 and 4 for capacitance and resistance, respectively. It proves that the presence of cholesterol contributes to the formation of a very stiff and virtually defect-free membrane that is significantly less permeable to ions and water than DMPC bilayer. The exposure of DMPC/Cholesterol bilayer to the 10 mM solution of melittin leads to the increase of the membrane permeability with the increasing time of melittin-lipid interactions (Fig. 5B). The trend of these changes is similar to that observed for DMPC bilayer exposed to 10 mM melittin. However, the time dependence is substantially different and longer exposure

Fig. 6. AFM images (1000  1000 nm2) of DMPC/Cholesterol (7:3) bilayer transferred onto the Au(111) electrode by combined Langmuir-Blodgett and Langmuir-Schaefer techniques. Image (A) shows intact DMPC bilayer, while image (B) presents the same sample after 120 minutes exposure to 10 mM melittin respectively. The average thickness of the film is shown in a bottom right corner of each image. The cross marks in panel (B) correspond to the points which are representative for the values of the thickness shown in the AFM image. The images were taken at the same spot on the sample and were recorded in a TRIS buffer (pH=7.4) at the potential of +0.26 V vs. Ag/AgCl.

60

J. Juhaniewicz, S. Sek / Electrochimica Acta 162 (2015) 53–61

is needed to observe increased permeability of the membrane. This shows clearly that DMPC/Cholesterol bilayer is more resistant to melittin action than the bilayer composed of pure DMPC. Initially the changes in capacitance with time are quite shallow reflecting gradual thinning of the membrane caused by its fluidization. After 180 minutes, the capacitance is doubled comparing with its value after 120 minutes. Such more pronounced increase may correspond to the abrupt change in dielectric properties of the film resulting from the formation of the pores which enable penetration of the membrane by ions and water. Charge transfer resistance follows similar trend, since the most pronounced drop of its value is observed between 120 and 180 minutes. Thus, the results obtained from impedance measurements seem to confirm the scenario where the action of melittin involves the fluidization of DMPC/Cholesterol membrane and further pore formation. The AFM studies revealed that indeed DMPC/Cholesterol bilayer immobilized on gold surface is compact and uniform. The image of intact membrane is presented in Fig. 6A. The thickness of the film determined from force-distance curves was 5.1  0.4 nm, which is in a very good agreement with the value reported by Chen and coworkers for the same systemwithin this potential range [49]. Upon the introduction of melittin, the changes in morphology of the bilayer were small during first 60 minutes. However, after 120 minutes, the difference in a structure was remarkable as demonstrated in Fig. 6B. The bilayer is not uniform and topographically higher islands are clearly visible. This is related to the coexistence of two phases. Brighter domains correspond to stiffer and more ordered phase protruding above the fluid-like phase represented by darker regions. The thickness of the bilayer measured at the location represented by the red mark in Fig. 6B is 5.0  0.3 nm, while for the white mark it is 4.0  0.3 nm. This confirms coexistence of two different states of the bilayer. Moreover, the phase imaging also reveals some differences in energy dissipation during the tip oscillation (see Figs. 4C and 4D). Initially, the bilayer is uniform in terms of viscoelastic properties. However, inhomogeneity is clearly visible and the phase shift becomes larger after 120 minutes of exposure to melittin. Thus the AFM imaging demonstrates that melittin action leads to local disordering of the bilayer and partial melting of the membrane. This may be considered as a crucial step for the reorientation of the peptide within the membrane and further formation of the pores. It is evident that the presence of cholesterol affects remarkably the mode of melittin action comparing with single component lipid bilayer. Cholesterol fills the space between DMPC acyl chains and influences the physical state of the membrane making it stiffer and less susceptible for the penetration by melittin molecules. Since it is buried in the chain region of the bilayer, it may be reasonable to consider that chain packing in hydrophobic part provides significant contribution in establishing physical barrier for peptide insertion. 4. CONCLUSIONS We have demonstrated that electrochemical methods supported by in situ AFM imaging enable the evaluation of the mechanisms involved in melittin membranolytic activity. In case of single component DMPC bilayer, the action of melittin at 10 mM concentration leads to quick degradation of the membrane, which undergoes micellization and subsequent partial dissolution. These results indicate that under such conditions melittin acts according to detergent-like mechanism. The mode of melittin action is substantially different when the peptide concentration is lowered down to 1 mM. As demonstrated by AFM imaging, the changes in DMPC bilayer morphology are less rapid and initially peptide adsorbs on top of the membrane. This observation is also supported by electrochemical results, which indicate that the permeability of the membrane is decreased. Longer exposure to melittin leads to the

fluidization of the DMPC film, which suggests that adsorbed peptide molecules disturb the arrangement of the lipid headgroups within the polar region and further cause disordering of the membrane, most likely by inducing the tilting of the hydrophobic chains. This process is accompanied by the reorientation and insertion of the peptide into the bilayer, which finally leads to the formation of large defects and numerous pinholes. This shows that clear differentiation between carpet and toroidal pore mechanism is difficult here, since both can be ascribed to the observed mode of melittin action. Thus, it cannot be ruled out that under such conditions melittin acts according to mixed mechanism. What is interesting, yet different behavior of melittin was observed for DMPC/Cholesterol bilayer. In this case, the susceptibility of the membrane to melittin action was reduced and the overall blocking properties of the lipid film were much better comparing with single component DMPC bilayer. First of all, the step involving adsorption of the peptide on top the bilayer was not observed. Thus, we conclude that melittin percolates into the polar head region immediately after approaching the bilayer outer plane. The difference in behavior comparing with single component DMPC bilayer is related to the fact that polar heads of lipid molecules are less densely packed in the presence of cholesterol. This facilitates the penetration of the polar region. However, the hydrophobic part of the bilayer seems to be much more resistant to melittin action and even after one hour of exposure, the diffusion of the redox probe to electrode surface is still negligible. Thus, we conclude that it is hydrocarbon chain region of the bilayer that is responsible for reduced melittin activity. The role of the cholesterol is twofold in this case. First of all, cholesterol increases packing density of the hydrocarbon chains within the membrane and leads to the increase of its elastic modulus [30,50]. In other words, the bilayer is tightened within its hydrophobic region. On top of that, cholesterol molecules may interact with tryptophan residues of melittin and such interaction simply slows down the penetration as demonstrated by Raghuraman and Chattopadhyay [51]. This may explain why the membrane withstands high concentration of melittin and as a consequence extended time is needed to fluidize the bilayer. Once it is achieved, the diffusion controlled transport through the bilayer is possible. At this point, we would like to emphasize that the fluidization of the membrane by melittin seems to be a common feature for all solid supported lipid bilayers studied. Therefore, it is reasonable to assume that it is a crucial step for melittin action which enables the peptide molecules to adopt proper orientation either for pore formation or disruption of the membrane. ACKNOWLEDGMENTS This paper is dedicated to prof. Jacek Lipkowski on the occasion of his 70th birthday. Financial support for the research was provided by National Science Centre (NCN Grant Opus No. 2012/05/ B/ST4/01243). The study was carried out at the Biological and Chemical Research Centre, University of Warsaw, established within the project co-financed by European Union from the European Regional Development Fund under the Operational Programme Innovative Economy, 2007–2013. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.electacta.2014.10.039. References [1] R.E.W. Hancock, Expert Opinion on Investigational Drugs 9 (2000) 1723. [2] Y. Li, Q. Xiang, Q. Zhang, Y. Huang, Z. Su, Peptides 37 (2012) 207. [3] J.P.S. Powers, R.E.W. Hancock, Peptides 24 (2003) 1681.

J. Juhaniewicz, S. Sek / Electrochimica Acta 162 (2015) 53–61 [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28]

Y. Shai, Biochimica et Biophysica Acta 1462 (1999) 55. Y. Shai, Biopolymers 66 (2002) 236. H. Sato, J.B. Feix, Biochimica et Biophysica Acta 1758 (2006) 1245. K. Matsuzaki, K. Sugishita, M. Harada, N. Fuji, K. Miyajima, Biochimica et Biophysica Acta 1327 (1997) 119. M. Wu, E. Maier, R. Benz, R.E. Hancock, Biochemistry 38 (1999) 7235. T.C. Terwilliger, D. Eisenberg, Journal of Biological Chemistry 257 (1982) 6010. T. Katsu, M. Kuroko, T. Morikawa, K. Sanchika, Y. Fujita, H. Yamamura, M. Uda, Biochimica et Biophysica Acta–Biomembranes 983 (1989) 135. C.E. Dempsey, Biochimica et Biophysica Acta–Reviews on Biomembranes 1031 (1990) 143. H. Raghuraman, A. Chattopadhyay, Bioscience Reports 27 (2007) 189. E. Sackmann, Science 271 (1996) 43. L.J. Johnston, Langmuir 23 (2007) 5886. M. Tanaka, E. Sackmann, Nature 437 (2005) 656. P.S. Cremer, S.G. Boxer, Journal of Physical Chemistry B 103 (1999) 2554. H.M. Seeger, A. Di Cerbo, A. Alessandrini, P. Facci, Journal of Physical Chemistry B 114 (2010) 8926. Z.V. Leonenko, A. Carnini, D.T. Cramb, Biochimica et Biophycisa Acta 1509 (2000) 134. J. Lipkowski, Physical Chemistry Chemical Physics 12 (2010) 13853. C. Steinem, H.J. Galla, A. Janshoff, Physical Chemistry Chemical Physics 2 (2000) 4580. L. Becucci, M. Papini, D. Mullen, A. Scaloni, G. Veglia, R. Guidelli, Biochimica et Biophysica Acta–Biomembranes 1808 (2011) 2745. W.K. Chang, W.C. Wimley, P.C. Searson, K. Hristova, M. Merzlyakov, Biochiomica et Biophysica Acta 1778 (2008) 2430. A. Plant, Langmuir 9 (1993) 2764. R.P. Janek, W.R. Fawcett, A. Ulman, Langmuir 14 (1998) 3011. X. Bin, I. Zawisza, J.D. Goddard, J. Lipkowski, Langmuir 21 (2005) 330. I. Zawisza, X. Bin, J. Lipkowski, Bioelectrochemistry 63 (2004) 137. M. Li, M. Chen, E. Sheepwash, C.L. Brosseau, H. Li, B. Pettinger, H. Gruler, J. Lipkowski, Langmuir 24 (2008) 10313. I. Burgess, S.L. Horswell, G. Szymanski, J. Lipkowski, J. Majewski, S. Satija, Biophysical Journal 86 (2004) 1763.

61

[29] R. Lipowski, E. Sackmann, Eds. Structure and Dynamics of Membranes, Elsevier, Amsterdam, 1995. [30] A. Kessel, N. Ben-Tal, S. May, Biophysical Journal 81 (2001) 643. [31] T. Benachir, M. Monette, J. Grenier, M. Lafleur, European Biophysics Journal 25 (1997) 201. [32] P. Wessman, A.A. Strömstedt, M. Malmsten, K. Edwards, Biophysical Journal 95 (2008) 4324. [33] L. Becucci, R.R. Leon, M.R. Moncelli, P. Rovero, R. Guidelli, Langmuir 22 (2006) 6644. [34] A.M. Feigin, J.H. Teeter, J.G. Brand, Biochemical and Biophysical Reserach Communications 211 (1995) 312. [35] J. Clavilier, Journal of Electroanalytical Chemistry 107 (1980) 211. [36] D. Marsh, Biochimica et Biophysica Acta 1286 (1996) 183. [37] R.A. Demel, W.S. Geurts van Kessel, R.F. Zwaal, B. Roelofsen, L.L. van Deenen, Biochimica et Biophysica Acta 406 (1975) 97. [38] A.S. Bondarenko, G.A. Ragoisha, Progress in Chemometrics Research, Nova Science Publishers, 2005, pp. 89. [39] S.C. Quay, C.C. Condie, Biochemistry 22 (1983) 695. [40] K. Hall, T.H. Lee, M.I. Aguilar, Journal of Molecular Recognition 24 (2011) 108. [41] J.P. Pratt, D.J. Ravnic, H.T. Huss, X. Jiang, B.S. Orozco, S.J. Mentzer, In Vitro Cellular & Developmental Biology, Animal 41 (2005) 349. [42] H.J. Butt, B. Capella, M. Kappl, Surface Science Reports 59 (2005) 1. [43] Y.F. Dufrene, T. Boland, J.W. Schneider, W.R. Barger, G.U. Lee, Faraday Discussions 111 (1998) 79. [44] D.A. Brevnov, H.O. Finklea, Langmuir 16 (2000) 5973. [45] R. Fettiplace, D.M. Andrews, D.A. Haydon, Journal of Membrane Biology 5 (1971) 277. [46] J.F. Popplewell, M.J. Swann, N.J. Freeman, C. McDonnell, R.C. Ford, Biochimica et Biophysica Acta - Biomembranes 1768 (2007) 13. [47] B. Bechinger, K. Lohner, Biochimica et Biophysica Acta 1758 (2006) 1529. [48] X. Bin, J. Lipkowski, Journal of Physical Chemistry B 110 (2006) 26430. [49] M. Chen, M. Li, C.L. Brosseau, J. Lipkowski, Langmuir 25 (2009) 1028. [50] T. Hianik, M. Haburcak, General Physiology and Biophysics 12 (1993) 283. [51] H. Raghuraman, A. Chattopadhyay, Biophysical Journal 87 (2004) 2419.

Suggest Documents