Direct Observation of Nanometer-Scale Pores of Melittin in Supported

Article pubs.acs.org/Langmuir

Direct Observation of Nanometer-Scale Pores of Melittin in Supported Lipid Monolayers Diana Giménez, Orlando L. Sánchez-Muñoz, and Jesús Salgado* Institute of Molecular Science (ICMol), University of Valencia. C/Catedrático José Beltrán, 2, 46980 Paterna, Valencia, Spain S Supporting Information *

ABSTRACT: Melittin is the most studied membrane-active peptide and archetype within a large and diverse group of pore formers. However, the molecular characteristics of melittin pores remain largely unknown. Herein, we show by atomic force microscopy (AFM) that lipid monolayers in the presence of melittin are decorated with numerous regularly shaped circular pores that can be distinguished from nonspecific monolayer defects. The specificity of these pores is reinforced through a statistical evaluation of depressions found in Langmuir−Blodgett monolayers in the presence and absence of melittin, which eventually allows characterization of the melittininduced pores at a quantitative low-resolution level. We observed that the large majority of pores exhibit near-circular symmetry and a Gaussian distribution in size, with a mean diameter of ∼8.7 nm. A distinctive feature is a ring of material found around the pores, made by, on average, three positive peaks, with a height over the level of the lipidic background of ∼0.23 nm. This protruding rim is most likely due to the presence of melittin near the pore border. Although the current resolution of the AFM images in the {x, y} plane does not allow distinction of the specific organization of the peptide molecules, these results provide an unprecedented view of melittin pores formed in lipidic interfaces and open new perspectives for future structural investigations of these and other pore-forming peptides and proteins using supported monolayers.

■

simulations.13,25,26 This helix is partially buried ∼1.7 nm from the bilayer center, at the level of the glycerol groups of phospholipids,20 where it binds almost parallel to the membrane plane, in natural correspondence with its amphipathic character. The orientation of melittin is further modulated by its concentration in the membrane, and by means of a variety of methods, at least two orientational states have been described for this peptide.8,12,23,27 From a series of experiments performed by the group of Huang, including equilibrium structural studies by circular dichroism and X-ray diffraction,12 and simultaneous kinetic measurements of area expansion and leakage in single giant vesicles,8,28,29 the two orientation states of melittin have been assigned to a planarly bound S-state and a perpendicularly inserted I-state.30,31 The Sstate corresponds to the natural absorption of the amphipathic helices in the interfacial region. This binding mode is a source of membrane tension, as can be inferred from measurements of membrane thinning31 and area expansion.8,28,29 At a limiting tension, reached at a threshold peptide-to-lipid molar ratio, the membrane yields, and pores are formed. This is coupled to the change in orientation of some melittin molecules, which then convert to the I-state.

INTRODUCTION Melittin, a major component of bee venom,1 is a paradigmatic membrane-active peptide that has been studied extensively because of its intrinsic antimicrobial and anticancer activities.2−4 It works by increasing the permeability of lipid membranes, and it is thus used as a model for the general understanding of ion channels and pores5−9 and as a scaffold for the design of new pore-forming peptides.10 However, to date, the structure of melittin bound to membranes, the molecular characteristics of the “active” pore state, and the mechanism of pore formation remain unknown and/or incomplete.6,8,11−17 A particular aspect under dispute is the very existence of pores and their relevance for the activity of melittin and similar peptides.8,15 This controversy is important because melittin is a key example and archetype within the vast family of membrane-active, pore-forming peptides,7,18 which are of growing biomedical and biotechnological interest. It is widely accepted that melittin folds as an α-helix when it binds to lipid interfaces.6,17,19−24 A high-resolution structure of melittin in aqueous environments was solved early by X-ray crystallography, where the peptide forms bent helices of strong amphipathic character that are assembled as tetramers.11,19 Although this structure does not correspond to a membranebound state, it was promptly reasoned that the amphipathicity of the monomer helices can explain binding to the membrane interface.19,20,24 Bent helical melittin has indeed been described in the membrane-bound state, both in experiments6 and in © XXXX American Chemical Society

Received: October 30, 2014 Revised: January 30, 2015

A

DOI: 10.1021/la504293q Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

702BAM, 729.2-cm2 total area). Filter paper balances were used for all experiments (Whatman CHR1, 10-mm width, 20.6-mm perimeter). Peptide-only monolayers were prepared using aliquots of the melittin stock solution in DMSO, filtered through polytetrafluoroethylene (PTFE) membranes (0.4- μm pore size, Sartorius Stedim Biotech GmbH). These were injected with a Hamilton microsyringe into a continuously stirred subphase [tris(hydroxymethyl)aminomethane (Tris) 25 mM, NaCl 50 mM, pH 7.4] in the Langmuir trough, to reach final concentrations ranging from 0 to 1 μM. Control experiments in the absence of peptide were carried out by subphase injection of DMSO up to 0.1% v/v. The changes in surface pressure due to peptide adsorption were continuously monitored as a function of time, until reaching maximum stable values (πmax). For lipid-only monolayers, stock solutions of dipalmitoylphosphatidylcholine (DPPC; Avanti Polar lipids) were prepared fresh by dissolving the appropriate amount of lipid in 3:1 CHCl3/methanol mixtures. These solutions were filtered, subsequently spread gently onto the air/buffer interface [2-[4-(2-hydroxyethyl)-1-piperazinyl]ethanesulfonic acid (HEPES), 25 mM, pH 7.4] using a Hamilton microsyringe, and allowed to stabilize for 30 min to achieve solvent evaporation. After that, isotherms were recorded continuously under symmetric compression, at a constant barrier speed of 3 cm2/min, using NIMA 542 control software. The same procedure was followed for preparing codispersed monolayers of DPPC and melittin, with the only difference being that the lipid and peptide components were codissolved in methanol before being spread at the air/buffer interface. For mixed lipid/peptide monolayers at constant area, lipid-only monolayers were prepared first as explained above, to reach a final surface pressure of 12.5 mN/m. Prior to melittin addition, the monolayers were allowed to stabilize for additional 30 min to ensure a reproducible final constant area under stirring conditions. On the other hand, for control experiments, patches of these monolayers were transferred onto glass solid supports (see below) before peptide addition. Then, melittin was added to the desired concentration by injection into the continuously stirred subphase of freshly made and filtered peptide solutions in DMSO (same conditions as explained above). The changes in surface pressure due to peptide adsorption in the monolayers were monitored continuously as a function of time, until reaching maximum stable values (πmax). In control experiments, DMSO (0.1% v/v) was added under the same conditions. For mixed lipid/peptide monolayers at constant surface pressure, different aliquots of the lipid stock solution in 3:1 CHCl3/methanol were spread to reach a final surface pressure of 12.5, 22.5, or 30 mN/m and allowed for solvent evaporation (30 min) at constant surface pressure. For controls, patches of these monolayers were transferred to solid supports (see below) before peptide addition. On the other hand, because the relaxation times of the monolayers depended on the particular surface pressures, in all cases, the monolayers were allowed to equilibrate for at least 100 min before transfer to supports or addition of peptide. Melittin was added to the desired concentration by injection into the continuously stirred subphase of freshly made and filtered peptide solutions in DMSO (same conditions as explained above). The changes in surface area due to peptide adsorption in the monolayers were monitored continuously as a function of time until reaching maximum stable values. In cases of very slowly increasing area, due to slow peptide binding, we took the area corresponding to 200 min of equilibration time. In control experiments, DMSO (0.1% v/v) was added under the same conditions. Supported Monolayers. For preparation of supported monolayers, we employed the Langmuir−Blodgett (LB) method41 using a rectangular Teflon KSV Minimicro-2 LB system (KSV Instruments Ltd., Helsinki, Finland) with an internal area of 8720 mm2 (51 mm × 170.98 mm) and a depth of 0.50 mm. The surface pressure was monitored with a precision of 0.1 mN/m by the Wilhelmy plate method. Equilibrated lipid or lipid−peptide monolayers, prepared as described above for each of the assayed conditions, were transferred onto glass coverslips (VWR); cleaned with alkaline detergent; rinsed with Milli-Q water; thoroughly washed with isopropanol (three times), methanol (three times), and ethanol (three times); and finally dried under a stream of nitrogen. For transfer of monolayers using the LB

The pores are viewed as the response of the bilayer to excess interfacial area, due to melittin binding.8 They are stable at least in preparations of supported lipid bilayers, where these pores have been measured by neutron scattering12,32 and their lipidic contour has been determined using grazing-angle X-ray anomalous diffraction.8 Such a lipidic structure of melittininduced pores has the distinctive characteristics of a lipid pore with a toroidal shape, with the two interfaces of monolayers fused at the pore wall.12 The existence of discrete pores induced by melittin can also be inferred from studies of leakage kinetics in single giant vesicles using fluorescence microscopy8,29,33 and from a combined elipsometry and laser scanning microscopy analysis in supported bilayers.34 The pore structure and mechanism have been modeled by molecular dynamics simulations,14,35 although the latter type of studies cannot yet provide realistic descriptions of peptide−membrane complexes.36 Therefore, the configuration of I-state melittin molecules, associated with pores, has thus far resisted characterization. This active state of melittin is assumed to be bound to the pore rim, where it has been given the role of increasing the pore stability by reducing the line tension.31 Importantly, although the pores likely originate and stabilize as a result of the concerted action of various melittin molecules, they apparently do not require the existence of direct peptide− peptide interactions, and in fact, the presence of melittin oligomers in the bilayer environment has repeatedly been discarded.24,37 In this work, we report the direct observation of melittininduced pores by atomic force microscopy (AFM) in supported lipid monolayers. Because the interaction of melittin with lipid bilayers occurs essentially at the interface level, the monolayer provides an analogous docking site for the peptide. The monolayer system is also adequate for trapping relevant structures of peptide−membrane complexes, including pores, whose molecular details can then be investigated using nanoscopic methods, such as AFM38 or scanning tunneling microscopy (STM).39,40 From the statistical analysis of AFM topographic images, we describe the nanometer-scale properties of the melittin pores, including a rim of protruding material that is likely due to melittin molecules.

■

MATERIALS AND METHODS

Peptide Synthesis and Purification. Melittin, with the sequence NH3+−GIGAVLKVLTTGLPALISWIKRKRQQ−CONH2, was made by solid-phase synthesis using FastMoc chemistry in an ABI-433 automatic peptide synthesizer (Applied Biosystems). Rink-AmideChemMatrix resin (Iris Biotech GmbH) was used as a solid support, and the peptide chain was elongated by a double coupling reaction for each Fmoc-amino acid (from Sigma) to optimize the yield. The peptide was deprotected at the N-terminal and side-chain groups, cleaved from the resin, and purified by reverse-phase high-performance liquid chromatography (HPLC; XBridge BEH3000 C18 5 μM, 10 × 250 mm column, Waters) on a Waters 2695 Separation Module. The fractions containing melittin, confirmed by matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry, with a purity >95% according to analytical HPLC, were stored lyophilized. Stock solutions for use in subsequent experiments were prepared in dimethyl sulfoxide (DMSO), with peptide concentrations measured by absorbance at 280 nm. Preparation and Study of Monolayer Films. All monolayer experiments were performed at least twice at 22 °C inside a thermostated, dust-free room. For isotherms of surface pressure versus mean molecular area, a rectangular NIMA Teflon-coated trough was used, equipped with Teflon barriers and a Wilhelmy balance (type B

DOI: 10.1021/la504293q Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Figure 1. Adsorption of melittin to DPPC monolayer films. (A) Kinetic profiles showing the change in surface pressure (π) with time for additions of the peptide in the subphase at different concentrations. The values of concentration are indicated in the graph and refer to the bulk amounts of peptide initially added by injection into the subphase. For a better comparison with other studies, a 75 nM melittin concentration in our samples corresponds to an estimated P/L ratio of 0.14. In all cases, prior to peptide addition, the lipid monolayer was equilibrated at πi = 12.5 mN/m. A control trace with no peptide is included, in which we added DMSO (the carrier solvent for peptide additions) up to 0.1% v/v. (B) Increments of surface pressure (Δπ) due to the presence of peptide (at the concentrations indicated in the graph) measured 3.2 h after peptide injection. Because the addition of DMSO to the control monolayers produced a smaller decrease of surface pressure, for the samples with melittin, the values of Δπ were calculated with respect to the average equilibrium pressure in the controls. The error bars represent the standard deviations (±SD). technique, the glass slides were immersed vertically into the trough dipping well at a constant speed of 1 cm/min, equilibrated for 10 min, and removed again vertically at the same constant speed. This allowed sections of the monolayer with an area of 8 × 18 mm2 to be transferred onto each side of the solid support. Epifluorescence Microscopy. Epifluorescence microscopy was carried out using a Nikon Eclipse Ti−S inverted microscope equipped with Nikon FITC and UV filters. Images were recorded at 1280 × 1024 resolution in no-binning mode using a high-sensitivity cooled CCD Nikon DS-Qi1 camera and the NIS-Elements software. An achromat LWD DL 40× Ph2 objective was used in all cases. The raw gray-scale images were pseudocolored using green for the FITC channel and red for UV channel. A minimum of 15 images corresponding to both sides of the glass support were recorded. Because the number and shape of observed domains were variable, even for same conditions and within the same day and same stock solutions, in cases of mixed lipid/peptide monolayers, for comparison, we also transferred and analyzed the same monolayers before peptide injection. These control, lipid-only monolayers were considered as zero-time states of the equilibration process in the presence of peptide. In cases of domain coexistence, the fractional area corresponding to each type of domain was calculated as an average of the values from a minimum of 15 images taken from each transferred monolayer, using the Image-J software package (imagej.nih.gov/). AFM Imaging. The nanometer-scale analysis of glass-supported LB monolayers was made by AFM in air and in tapping mode with a NanoScope IVa Controller from Veeco. We used high-resolution silicon probes of the type PPP-NCH from Nanosensors. These tips are shaped as polygonal pyramids with a height of 10−15 mm, a half-cone angle at the tip apex of 10°, and a tip radius of 0 having its maximum height (h) within a radial distance from the depression center of 4rp, where rp is the radius of a disk with an area equal to Aabs. The same strategy of analysis was applied for LB lipid monolayers in the presence (37.5 nM) and in the absence of melittin, for the different assayed conditions, which allowed comparison among them. The numbers of considered features along the selection procedure (Nt, Nd, and Np) and their average areas are given in the Supporting Information in Table S1 (in the presence of melittin) and Table S2 (controls, in the absence of melittin). A summary of the characteristics of selected melittin pores is given in Table 1. Interestingly, with no exception, all monolayers with melittin showed a good number of pore features, which, in most cases (particularly for insertion at constant area and cospreading), corresponded to a large proportion of all depressions found in the monolayer

observed in the presence of melittin are qualitatively distinguishable from defects found in LB lipid-only monolayers in a number of characteristics, including their abundance, size, shape, and rim made of protruding material. To provide statistical support to the specificity of melittin-induced pores and to better describe their properties at a quantitative level, we performed a systematic unbiased analysis of the images. We first note that the size and geometry of the AFM probe imposes limitations for the characterization of the topography of analyzed surfaces. In our specific case, the probe was thick compared to the size of pores and most defects carved in the monolayers, and consequently, the probe tip could not easily reach the bottom level of these features. The geometry of the tip may also produce overhanging artifacts that may limit the resolution in the {x, y} plane as the probe sampled the roughness of the surface, giving rounded edges, overestimating the area of protruding features, and underestimating the area of depressions. However, the AFM topographic analysis is very sensitive to the height of resolved positive peaks. Considering these limitations, to characterize the relevant topographic features systematically, the three-dimensional {x, y, z} matrices of flattened and calibrated AFM topographic images were conveniently represented as contour-level maps (see examples in Figure 7). Depressions in 0.5−2 μm2 monolayer areas were first localized as regions with a negative height (z < 0 nm). To these depressions, with a total number of Nt, we applied the following two-step selection criteria (see Tables S1 and S2, Supporting Information): (1) First, we selected the depressions with a minimum depth of 0.5 nm and a minimum area enclosed by the 0.5-nm z level (Aabs) of 3.5 nm2. This reduced the number of cases to Nd. (2) Second, of the Nd depressions, we selected those that loosely approximated a circular shape according to the following strategy: The center of the area encompassed by the 0.5-nm level of each depression and the distances from this center to each point of the 0.5-nm level were determined. We then chose the depressions that simultaneously met the criteria

Table 1. Characteristics of Pores Found in Lipid−Melittin Monolayers type of phase

Dp (pores/μm)

Aabsa (nm2)

nb

hc (nm2)

Subphase Injection at πi = 12.5 mN/m and Equilibration at Constant Area (πeq = 19.1 mN/m) LE + LC 26 64 ± 27 3.8 ± 1.4 0.3 ± 0.2 LE 24 49 ± 20 3.9 ± 1.3 0.3 ±0.20 LC 28 77 ± 27 3.6 ± 1.5 0.3 ± 0.2 Subphase Injection and Equilibration at Constant Pressure (πc = 30 mN/m) LC 16 40 ± 15 3.0 ± 1.3 0.2 ± 0.1 Cospreading at πeq = 12.5 mN/m LE + LC 35 65 ± 35 3.3 ± 2.1 0.3 ± 0.2 LE 25 51 ± 25 4.0 ± 0.7 0.4 ± 0.2 LC 47 79 ± 28 3.5 ± 2 0.3 ± 0.1 Cospreading at πeq = 30 mN/m LC 41 61 ± 32 2.7 ± 1.2 0.2 ± 0.1

1 ≤ (rmax /rmin) ≤ 2.5 1/1.5 ≤ (A abs /Acir ) ≤ 1.5

a Average area ± standard deviation. bAverage number of peaks present around a pore ± standard deviation. The peaks are defined as any region with z > 0 within a radial distance from the pore center of 4req, where req is the radius of a disk with an area equal to Aabs. c Average height of peaks present around a pore ± standard deviation.

where rmax and rmin are the maximum and minimum distances, respectively, from the center of the considered depression to points of its 0.5-nm height level and Acir is the area of a circle with a radius equal to (rmax + rmin)/2. I

DOI: 10.1021/la504293q Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Figure 8. Frequency distributions of parameters corresponding to the total number of pores found in supported monolayers in the presence of melittin. (A) Distribution of pore areas. (B) Distribution of the number of peaks present around a pore. (C) Distribution of the height of peaks present around pores. The total number of pores (Np from Table 1) was 226, and the analyzed monolayer area was 8.3 μm2. These pores were found in monolayers prepared under different conditions, as indicated by colors in the figure, but they group well in a single distribution, as represented by the histograms. Gaussian fits of the complete distributions are also shown (dashed red lines).

melittin-containing LB monolayers, which eliminates the possibility that the monolayer pores are primed by the glass substrate. On the other hand, the clear association of the pores with the presence of melittin and their seemingly regular characteristics (in size and shape) in LB monolayers prepared under different conditions and in LE and LC phases argue in favor of their link to melittin and strongly suggest that they are formed, or at least induced, by specific molecular organizations of this peptide. Although lipid monolayers alone (with 0.1% DMSO, which is the carrier solvent for peptide addition) were sometimes carved by depressions, in our controls, most of these features observed in lipid-only samples were shallow and/or irregular and did not pass our two-step selection criteria to be considered as true pores (Table S2, Supporting Information). Only for one of the controls, namely, LB monolayers prepared at an equilibrium pressure of 30 mN/m, did a small proportion of the observed depressions (∼16%) fulfill the criteria for classification as pores (Figure 6G−I). However, the “pores” found in this lipid-only case differed from true pores seen in the presence of melittin in the distributions and mean values of Aabs, n, and h. They also lacked a clear protruding ring around the pore, observable in topography and phase images (Figure 6H). Thus, our unbiased analysis provides a ground for distinguishing nonspecific defects from melittin-induced pores found in LB monolayers. Regarding the significance of these pores, one should consider the particular characteristics and conditions of the samples used for this study. Lipid monolayer samples are a wellknown platform for preparing two-dimensional crystals of proteins.52 For the case of pore-forming peptides, LB monolayers have been used to investigate gramicidin A by AFM38 and STM39 and alamethicin by STM.40 In all of these cases, the lipid monolayer provides a nativelike physicochemical environment and facilitates the formation of ordered arrays of self-assembled structures with close resemblance to the native structures. Although melittin manifests intrinsic interfacial activity at clean air/water interfaces (Figure S1, Supporting Information), it shows much stronger avidity for the interface of lipid monolayers (Figure 1), even when they are of zwitterionic character. That is, there is ample margin for studying specific effects of the interaction of this peptide with lipid monolayers, as well as specific structures formed in peptide/lipid-monolayer complexes. On the other hand, an important condition in our samples is low hydration, because the AFM measurements were made in air. This fact could be used to question the relevance of our observations for real, well-hydrated melittin in membranes.

(Table S1, Supporting Information). For the various cases with coexistence of phases, we observed no consistent difference in the number of pores or the values of n and h between phases, although in all cases with phase discrimination, the pore size (Aabs) was smaller in LE than in LC phases (Table 1). In contrast, among the control samples, without peptide, features that were classifiable as pores were observed only in monolayers transferred at πeq = 30 mN/m (Figure 6G−I and Table S2, Supporting Information). However, they were different from the pores found in the presence of melittin, as they corresponded to a small proportion of the total number of depressions seen in the monolayer, they had a very broad distribution of Aabs, they showed no peaks or only a few peaks near the pore (average n = 1.3 ± 1.2), and these peaks were small compared to peaks around the melittin pores (Table S2, Supporting Information). The characteristics of pores and differences between cases can be further appreciated in terms of the distributions of Aabs, n, and h. These distributions are similar to each other and can be fitted by Gaussian functions. The mean values of the distributions (fitted Gaussians) are given in Table 1. In the monolayers containing melittin, the mean values of Aabs vary between 40 ± 15 nm2 (for insertion at constant pressure, πc = 30 mN/m) and 79 ± 28 nm2 (for the LC phase in samples made by cospreading and equilibration at πeq = 12.5 mN/m). The overlap of all size distributions from melittin pores can be nicely fitted by a Gaussian, centered at Aabs = 59 ± 32 nm2(Figure 8A). The distributions of all n (Figure 8B) and h (Figure 8C) values are also fairly normal, all together supporting the conclusion that the pores observed in melittin-containing monolayers are robust, regular, and specific structures induced by the peptide.

■

DISCUSSION Cationic pore-forming peptides are attractive templates for promising applications in biomedicine and biotechnology,50,51 but new developments based on these systems are hindered by the lack of details about the molecular organization of their active structures in membranes. In this work, we report the first direct observation of melittin pores in a lipidic environment, by AFM analysis of supported LB monolayers. Specificity and Significance of Observed Pores. A first question in this study pertains to the specificity and significance of the observed pores. First, the background roughness of the glass plates used as supports (Figure S2, Supporting Information) does not seem to have any correspondence with the pattern, shape, and size of the pores observed in J

DOI: 10.1021/la504293q Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

reported in previous investigations of melittin8 and other similar peptides.54−56 Pore Size. As we have just discussed, the grouped distributions of pore properties can be taken to represent the general characteristics of melittin-induced pores, at least for a given lipid composition and bulk melittin concentration. In these results, one can expect only a minor influence of possible tip-broadening effects, which might at most affect the width of the distributions but not their mean values and Gaussian shape. For our assayed conditions, the global distribution of melittin pore areas (Figure 8) corresponds to a diameter of ∼8.7 nm (considering idealized cylindrical pores). This value is larger than the inside diameter of melittin pores in hydrated 1palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) and 1,2-didodecanoyl-sn-glycero-3-phosphocholine (DLPC) bilayers, measured by in-plane neutron scattering (d = 4.4 nm),12,32 but it is close to the largest of a group of values reported from a number of leakage studies (1−6 nm).12,17,57−59 On the other hand, the diameter of the lipidic profile of melittin-induced pores determined by grazing-angle X-ray diffraction is 0.7 nm, although this particularly small value has been attributed to the low hydration of the stacked bilayers used for the measurements.8 In any case, the larger pore size observed in this work can be ascribed to the use of monolayers. For example, even with comparable molecular organizations of pores in monolayer and bilayer samples, the smaller line tension in monolayers60 can give rise to pores of larger size. Taking all of these points into account, our observed pore size is not in disagreement with other data from the literature. Organization of Melittin Molecules. An important aspect is the distribution and organization of melittin molecules in the LB monolayers. However, the AFM images do not provide molecule-specific contrast. The topography images show only variations of height, from which one cannot easily infer the distribution of the peptides because the differences in brightness levels can also originate from lateral changes in the molecular organization of lipids, such as domains corresponding to different lipid phases. Additionally, because the peptide is partially inserted within the lipid interface region, the heights of peaks cannot be easily related to specific dimensions of the peptide molecule, such as the thickness and length expected for a melittin α-helix. On the other hand, the contrast in phase images can, in principle, be used to distinguish the peptides from the lipidic background, but this could as well originate from local variations of the conformation, packing, and dynamics of lipid chains. Despite these limitations, we suggest that the overall texture of the melittin-containing monolayers can be associated with the distribution of the peptide in the lipidic background, especially when there is overlap between topographic and phase features, as observed for most pores in the monolayers with melittin (Figure 5, columns of left and middle figures). Similarly, the presence in most cases of a protruding craterlike structure near the borders of pores suggests that some specific groupings of melittin molecules arrange and occupy positions surrounding these pores. The most likely assumption in all of these cases is that the peptiderich areas correspond to thickening of the monolayer with respect to the level of surrounding lipids. For surface-bound melittin, this could be due to the bulkiness of a helical bent structure inserted within the lipid interface of the monolayer.6,11,13,14,25,26 Using theoretical considerations, an increased thickening has been predicted near the edge of toroidal pores, due to curvature stress.31 In our images, a thickening of the

However, a significant amount of water should still be present in the polar, interfacial region of the supported monolayers, and this could be enough to allow for the proper structural organization of lipid and peptide molecules. In fact, lowhydration conditions have been also used in other structural studies of melittin pores, such as grazing-angle X-ray diffraction.8 Another aspect of the low hydration is that it could affect the chances of pore formation in the supported monolayer samples. Because we observed a relatively large number of pores (Dp = 16-47 pores/μm2, depending on conditions; Table 1), a likely possibility is that low hydration increases the probability of pore formation. Moreover, considering a model of tension-induced pores,18,31,53 it is also likely that the probability of pore formation in LB monolayers is higher than that in free-standing bilayers, in agreement with the relatively high density of pores observed here. In all of these respects, lipid monolayers are promising model systems for the structural study of pore-forming molecules in general. Properties of Melittin Pores. Melittin and Melittin Pores under Different Monolayer Conditions. In LB monolayers showing LE/LC phase coexistence, epifluorescence images indicate an apparent preference of melittin for the LE domains and/or deeper insertion of the Trp19 residue within the interface of these domains, compared to LC domains (Figures 2 and 3). Although both alternatives are in agreement with the higher compaction of LC domains, the overall decrease of Trp19 UV brightness at increased melittin concentration suggests that the UV patterns are mainly associated with different depths of binding or different orientations of the peptide helix with respect to the monolayer plane. This conclusion is in line with the distribution of melittin inferred from the density of melittin-induced pores (Dp) in the surface of monolayers, as observed in AFM images (Table 1). Thus, for a given type of monolayer, the Dp values are similar for LC and LE phases (as in monolayers prepared at constant area) or even larger in LC than in LE phases (case of cospreading and πeq = 12.5 mN/m). Different molecular organizations of melittin, depending on the monolayer compaction, could be responsible for the stabilization of phase-separated domains of lipids, which are visible at higher pressures in the presence of melittin than in the absence of this peptide (Figure 3C). This could also be the reason for our observation of smaller pores in LE phases (∼50 nm2) than in LC phases (∼78 nm2), in samples where the two phases coexist (Table 1). Nevertheless, our results show that the two lipid phases and corresponding modes of association of melittin are equally compatible with pore formation. On the other hand, despite a variety of sample conditions (monolayers prepared at constant area, constant high pressure, or cospreading of melittin and lipids), for a given melittin concentration (37.5 nM), the statistical analysis of AFM images (Table 1) allows for the description of a relatively homogeneous population of pores represented by the global distributions shown in Figure 8. These pores are characterized by their near-circular symmetry, a mean area of ∼59 ± 32 nm2, and a protruding rim with 3.2 ± 1.6 peaks and mean height of 0.23 ± 0.17 nm. The regularity of these distributions supports the robustness and specificity of the observed pore structures. It also agrees with the predictions made by Huang and coworkers on the basis of the analysis of pore energy in lipid bilayers.31 Additionally, our observation of well-defined specific pores formed, or induced, by melittin is an important fact that supports the existence of stable pores in membranes, as K

DOI: 10.1021/la504293q Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

(3) Raghuraman, H.; Chattopadhyay, A. Melittin: A MembraneActive Peptide with Diverse Functions. Biosci. Rep. 2007, 27, 189−223. (4) Dempsey, C. E. The Actions of Melittin on Membranes. Biochim. Biophys. Acta, Rev. Biomembr. 1990, 1031, 143−161. (5) Tosteson, M. T.; Tosteson, D. C. The Sting. Melittin Forms Channels in Lipid Bilayers. Biophys. J. 1981, 36, 109−116. (6) Vogel, H.; Jähnig, F. The Structure of Melittin in Membranes. Biophys. J. 1986, 50, 573−582. (7) Krauson, A. J.; He, J.; Wimley, W. C. Determining the Mechanism of Membrane Permeabilizing Peptides: Identification of Potent, Equilibrium Pore-Formers. Biochim. Biophys. Acta, Biomembr. 2012, 1818, 1625−1632. (8) Lee, M.-T.; Sun, T.-L.; Hung, W.-C.; Huang, H. W. Process of Inducing Pores in Membranes by Melittin. Proc. Natl. Acad. Sci. U.S.A. 2013, 110, 14243−14248. (9) Stephan, M.; Mey, I.; Steinem, C.; Janshoff, A. Combining Reflectometry and Fluorescence Microscopy: An Assay for the Investigation of Leakage Processes across Lipid Membranes. Anal. Chem. 2014, 86, 1366−1371. (10) Krauson, A. J.; He, J.; Wimley, W. C. Gain-of-Function Analogues of the Pore-Forming Peptide Melittin Selected by Orthogonal High-Throughput Screening. J. Am. Chem. Soc. 2012, 134, 12732−12741. (11) Terwilliger, T.; Eisenberg, D. The Structure of Melittin. II. Interpretation of the Structure. J. Biol. Chem. 1982, 257, 6016−6022. (12) Yang, L.; Harroun, T. A.; Weiss, T. M.; Ding, L.; Huang, H. W. Barrel-Stave Model or Toroidal Model? A Case Study on Melittin Pores. Biophys. J. 2001, 81, 1475−1485. (13) Irudayam, S. J.; Berkowitz, M. L. Binding and Reorientation of Melittin in a POPC Bilayer: Computer Simulations. Biochim. Biophys. Acta, Biomembr. 2012, 1818, 2975−2981. (14) Santo, K. P.; Irudayam, S. J.; Berkowitz, M. L. Melittin Creates Transient Pores in a Lipid Bilayer: Results from Computer Simulations. J. Phys. Chem. B 2013, 117, 5031−5042. (15) Wiedman, G.; Herman, K.; Searson, P.; Wimley, W. C.; Hristova, K. The Electrical Response of Bilayers to the Bee Venom Toxin Melittin: Evidence for Transient Bilayer Permeabilization. Biochim. Biophys. Acta, Biomembr. 2013, 1828, 1357−1364. (16) Wiedman, G.; Fuselier, T.; He, J.; Searson, P. C.; Hristova, K.; Wimley, W. C. Highly Efficient Macromolecule-Sized Poration of Lipid Bilayers by a Synthetically Evolved Peptide. J. Am. Chem. Soc. 2014, 136, 4724−4731. (17) Ladokhin, A. S.; White, S. H. ‘Detergent-Like’ Permeabilization of Anionic Lipid Vesicles by Melittin. Biochim. Biophys. Acta, Biomembr. 2001, 1514, 253−260. (18) Last, N. B.; Schlamadinger, D. E.; Miranker, A. D. A Common Landscape for Membrane-Active Peptides. Protein Sci. 2013, 22, 870− 882. (19) Terwilliger, T. C.; Weissman, L.; Eisenberg, D. The Structure of Melittin in the Form I Crystals and Its Implication for Melittin’s Lytic and Surface Activities. Biophys. J. 1982, 37, 353−361. (20) Hristova, K.; Dempsey, C. E.; White, S. H. Structure, Location, and Lipid Perturbations of Melittin at the Membrane Interface. Biophys. J. 2001, 80, 801−811. (21) Dempsey, C. E.; Butler, G. S. Helical Structure and Orientation of Melittin in Dispersed Phospholipid Membranes from Amide Exchange Analysis in Situ. Biochemistry 1992, 31, 11973−11977. (22) Dawson, C. R.; Drake, A. F.; Helliwell, J.; Hider, R. C. The Interaction of Bee Melittin with Lipid Bilayer Membranes. Biochim. Biophys. Acta, Biomembr. 1978, 510, 75−86. (23) Frey, S.; Tamm, L. K. Orientation of Melittin in Phospholipid Bilayers. A Polarized Attenuated Total Reflection Infrared Study. Biophys. J. 1991, 60, 922−930. (24) Schwarz, G.; Beschiaschvili, G. Thermodynamic and Kinetic Studies on the Association of Melittin with a Phospholipid Bilayer. Biochim. Biophys. Acta, Biomembr. 1989, 979, 82−90. (25) Andersson, M.; Ulmschneider, J. P.; Ulmschneider, M. B.; White, S. H. Conformational States of Melittin at a Bilayer Interface. Biophys. J. 2013, 104, L12−L14.

pore rim is absent or weak for purely lipidic pores (Figure 6), but there is a clear protrusion forming a ring near the border of most melittin-induced pores (Figure 5). This difference strongly supports the presence of melittin molecules near the pore border. Furthermore, the significant regularity in the numbers and heights of peaks forming the rims of the pores (Table 1 and Figure 8B,C) strongly suggests regularity in the structure of melittin and a specific arrangement of a particular number of peptides or groups of peptides. The number of peaks around a pore averages between 2.7 and 4.0 depending on the sample, with a height above the lipidic background of 0.2−0.4 nm. This might indicate a variable number of peptides forming a pore. However, because these peptides are partially immersed in the monolayers, some of them might not be detectable. In this respect, a value of four peptides or peptide groupings per pore seems to be a realistic number. All together, these facts suggest the existence of melittin oligomers, at least as components of the pore structures, although the present level of resolution does not allow a detailed description of the peptide arrangement. In conclusion, the presence of melittin in LB monolayers of DPPC induce the formation of multiple pores, distinguishable by their abundance and the regularity of their size and shape. These pores form similarly in LE and LC phases and in monolayers prepared under a variety of conditions. They display a disklike shape, a craterlike border that is probably made of melittin molecules, and a somewhat larger size compared to pores reported so far in bilayer samples. Although the specific molecular organization of melittin in the pores cannot be determined at the current resolution of the AFM images, these results are promising for future structural investigations of this and similar pore-forming systems.

■

ASSOCIATED CONTENT

S Supporting Information *

Intrinsic interfacial activity of melittin at air/water interfaces (Figure S1), control experiments corresponding to topography and phase images of a bare glass plate surface (Figure S2), characteristics of depressions found in mixed lipid/melittin monolayers and selection of those classified as pores (Table S1), and characteristics of depressions in control lipid-only monolayers and selection of those classified as pores (Table S2). This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Phone: +34963543016. Fax: +34963543576. E-mail: jesus. [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We acknowledge support from the Spanish MINECO (BFU2010-19118 and BFU2013-41648-P, financed in part by the European Regional Development Fund).

■

REFERENCES

(1) Habermann, E. Bee and Wasp Venoms. Science 1972, 177, 314− 322. (2) Gajski, G.; Garaj-Vrhovac, V. Melittin: A Lytic Peptide with Anticancer Properties. Environ. Toxicol. Pharmacol. 2013, 36, 697−705. L

DOI: 10.1021/la504293q Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir (26) Chen, C. H.; Wiedman, G.; Khan, A.; Ulmschneider, M. B. Absorption and Folding of Melittin onto Lipid Bilayer Membranes via Unbiased Atomic Detail Microsecond Molecular Dynamics Simulation. Biochim. Biophys. Acta, Biomembr. 2014, 1838, 2243−2249. (27) Chen, X.; Wang, J.; Boughton, A. P.; Kristalyn, C. B.; Chen, Z. Multiple Orientation of Melittin inside a Single Lipid Bilayer Determined by Combined Vibrational Spectroscopic Studies. J. Am. Chem. Soc. 2007, 129, 1420−1427. (28) Longo, M.; Waring, A.; Gordon, L.; Hammer, D. Area Expansion and Permeation of Phospholipid Membrane Bilayers by Influenza Fusion Peptides and Melittin. Langmuir 1998, 14, 2385− 2395. (29) Lee, M.; Hung, W.; Chen, F.; Huang, H. Mechanism and Kinetics of Pore Formation in Membranes by Water-Soluble Amphipathic Peptides. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 5087− 5092. (30) Huang, H. W. Action of Antimicrobial Peptides: Two-State Model. Biochemistry 2000, 39, 8347−8352. (31) Lee, M.-T.; Chen, F.-Y.; Huang, H. W. Energetics of Pore Formation Induced by Membrane Active Peptides. Biochemistry 2004, 43, 3590−3599. (32) He, K.; Ludtke, S. J.; Huang, H. W.; Worcester, D. L. Antimicrobial Peptide Pores in Membranes Detected by Neutron InPlane Scattering. Biochemistry 1995, 34, 15614−15618. (33) Kokot, G.; Mally, M.; Svetina, S. The Dynamics of MelittinInduced Membrane Permeability. Eur. Biophys. J. 2012, 41, 461−474. (34) Macháň, R.; Miszta, A.; Hermens, W.; Hof, M. Real-Time Monitoring of Melittin-Induced Pore and Tubule Formation from Supported Lipid Bilayers and Its Physiological Relevance. Chem. Phys. Lipids 2010, 163, 200−206. (35) Sengupta, D.; Leontiadou, H.; Mark, A. E.; Marrink, S.-J. Toroidal Pores Formed by Antimicrobial Peptides Show Significant Disorder. Biochim. Biophys. Acta, Biomembr. 2008, 1778, 2308−2317. (36) Wang, Y.; Zhao, T.; Wei, D.; Strandberg, E.; Ulrich, A. S.; Ulmschneider, J. P. How Reliable Are Molecular Dynamics Simulations of Membrane Active Antimicrobial Peptides? Biochim. Biophys. Acta, Biomembr. 2014, 1838, 2280−2288. (37) John, E.; Jähnig, F. Aggregation State of Melittin in Lipid Vesicle Membranes. Biophys. J. 1991, 60, 319−328. (38) Diociaiuti, M.; Bordi, F.; Motta, A.; Carosi, A.; Molinari, A.; Arancia, G.; Coluzza, C. Aggregation of Gramicidin A in Phospholipid Langmuir−Blodgett Monolayers. Biophys. J. 2002, 82, 3198−3206. (39) Sek, S.; Laredo, T.; Dutcher, J. R.; Lipkowski, J. Molecular Resolution Imaging of an Antibiotic Peptide in a Lipid Matrix. J. Am. Chem. Soc. 2009, 131, 6439−6444. (40) Pieta, P.; Mirza, J.; Lipkowski, J. Direct Visualization of the Alamethicin Pore Formed in a Planar Phospholipid Matrix. Proc. Natl. Acad. Sci. U.S.A. 2012, 109, 21223−21227. (41) Blodgett, K. B. Monomolecular Films of Fatty Acids on Glass. J. Am. Chem. Soc. 1934, 56, 495−495. (42) DeGrado, W. F.; Kézdy, F. J.; Kaiser, E. T. Design, Synthesis, and Characterization of a Cytotoxic Peptide with Melittin-Like Activity. J. Am. Chem. Soc. 1981, 103, 679−681. (43) Wackerbauer, G.; Weis, I.; Schwarz, G. Preferential Partitioning of Melittin into the Air/Water Interface: Structural and Thermodynamic Implications. Biophys. J. 1996, 71, 1422−1427. (44) Birdi, K. S. Self-Assembly Monolayer Structures of Lipids and Macromolecules at Interfaces; Springer: New York, 1999. (45) Maget-Dana, R. The Monolayer Technique: A Potent Tool for Studying the Interfacial Properties of Antimicrobial and MembraneLytic Peptides and Their Interactions with Lipid Membranes. Biochim. Biophys. Acta 1999, 1462, 109−140. (46) Lad, M. D.; Birembaut, F.; Clifton, L. A.; Frazier, R. A.; Webster, J. R. P.; Green, R. J. Antimicrobial Peptide−Lipid Binding Interactions and Binding Selectivity. Biophys. J. 2007, 92, 3575−3586. (47) Flach, C. R.; Prendergast, F. G.; Mendelsohn, R. Infrared Reflection-Absorption of Melittin Interaction with Phospholipid Monolayers at the Air/Water Interface. Biophys. J. 1996, 70, 539−546.

(48) Tsukanova, V.; Grainger, D. W.; Salesse, C. Monolayer Behavior of NBD-Labeled Phospholipids at the Air/Water Interface. Langmuir 2002, 18, 5539−5550. (49) Nag, K.; Perez-Gil, J.; Ruano, M. L. F.; Worthman, L. A. D.; Stewart, J.; Casals, C.; Keough, K. M. W. Phase Transitions in Films of Lung Surfactant at the Air−Water Interface. Biophys. J. 1998, 74, 2983−2995. (50) Bayley, H.; Jayasinghe, L. Functional Engineered Channels and Pores. Mol. Membr. Biol. 2004, 21, 209−220. (51) Majd, S.; Yusko, E. C.; Billeh, Y. N.; Macrae, M. X.; Yang, J.; Mayer, M. Applications of Biological Pores in Nanomedicine, Sensing, and Nanoelectronics. Curr. Opin. Biotechnol. 2010, 21, 439−476. (52) Ribi, H. O.; Ludwig, D. S.; Mercer, K. L.; Schoolnik, G. K.; Kornberg, R. D. Three-Dimensional Structure of Cholera Toxin Penetrating a Lipid Membrane. Science 1988, 239, 1272−1276. (53) Fuertes, G.; Giménez, D.; Esteban-Martín, S.; Sánchez-Muñoz, O. L.; Salgado, J. A Lipocentric View of Peptide-Induced Pores. Eur. Biophys. J. 2011, 40, 399−415. (54) Tamba, Y.; Ariyama, H.; Levadny, V.; Yamazaki, M. Kinetic Pathway of Antimicrobial Peptide Magainin 2-Induced Pore Formation in Lipid Membranes. J. Phys. Chem. B 2010, 114, 12018− 12026. (55) Fuertes, G.; García-Sáez, A. J.; Esteban-Martín, S.; Giménez, D.; Sánchez-Muñoz, O. L.; Schwille, P.; Salgado, J. Pores Formed by Baxα5 Relax to a Smaller Size and Keep at Equilibrium. Biophys. J. 2010, 99, 2917−2925. (56) Last, N. B.; Miranker, A. D. Common Mechanism Unites Membrane Poration by Amyloid and Antimicrobial Peptides. Proc. Natl. Acad. Sci. U.S.A. 2013, 110, 6382−6387. (57) Rex, S. Pore Formation Induced by the Peptide Melittin in Different Lipid Vesicle Membranes. Biophys. Chem. 1996, 58, 75−85. (58) Ladokhin, A. S.; Selsted, M. E.; White, S. H. Sizing Membrane Pores in Lipid Vesicles by Leakage of Co-Encapsulated Markers: Pore Formation by Melittin. Biophys. J. 1997, 72, 1762−1766. (59) Matsuzaki, K.; Yoneyama, S.; Miyajima, K. Pore Formation and Translocation of Melittin. Biophys. J. 1997, 73, 831−838. (60) Sriram, I.; Schwartz, D. K. Line Tension between Coexisting Phases in Monolayers and Bilayers of Amphiphilic Molecules. Surf. Sci. Rep. 2012, 67, 143−159.

M

DOI: 10.1021/la504293q Langmuir XXXX, XXX, XXX−XXX