Permeabilization Assay for Antimicrobial Peptides Based on PoreSpanning Lipid Membranes on Nanoporous Alumina Henrik Neubacher,† Ingo Mey,† Christian Carnarius, Thomas D. Lazzara, and Claudia Steinem* Institute of Organic and Biomolecular Chemistry, University of Göfttingen, Tammannstraße 2, 37077 Göttingen, Germany S Supporting Information *

ABSTRACT: Screening tools to study antimicrobial peptides (AMPs) with the aim to optimize therapeutic delivery vectors require automated and parallelized sampling based on chip technology. Here, we present the development of a chip-based assay that allows for the investigation of the action of AMPs on planar lipid membranes in a time-resolved manner by fluorescence readout. Anodic aluminum oxide (AAO) composed of cylindrical pores with a diameter of 70 nm and a thickness of up to 10 μm was used as a support to generate pore-spanning lipid bilayers from giant unilamellar vesicle spreading, which resulted in large continuous membrane patches sealing the pores. Because AAO is optically transparent, fluid single lipid bilayers and the underlying pore cavities can be readily observed by three-dimensional confocal laser scanning microscopy (CLSM). To assay the membrane permeabilizing activity of the AMPs, the translocation of the water-soluble dyes into the AAO cavities and the fluorescence of the sulforhodamine 101 1,2-dihexadecanoyl-sn-glycero-3-phosphoethanol-l-amine triethylammonium salt (Texas Red DHPE)labeled lipid membrane were observed by CLSM in a time-resolved manner as a function of the AMP concentration. The effect of two different AMPs, magainin-2 and melittin, was investigated, showing that the concentrations required for membrane permeabilization and the kinetics of the dye entrance differ significantly. Our results are discussed in light of the proposed permeabilization models of the two AMPs. The presented data demonstrate the potential of this setup for the development of an on-chip screening platform for AMPs.

INTRODUCTION Chip-based membrane assays are of high relevance not only to study transmembrane proteins but also to investigate small, membrane-active molecules.1,2 Biologically active molecules, such as antimicrobial peptides (AMPs), have been shown to be able to interact with lipid membranes, can form pores within lipid bilayers, and translocate large cargos across the latter.3−5 AMPs are discussed as alternatives to common antibiotics,6,7 fueled by the advent of methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant S. aureus (VRSA) that are known as fast-growing bacteria, which are resistant to current available antibiotics.8 They are found in a variety of different species to inhibit or regulate microbial growth as part of the innate immune response.9−11 Almost all AMPs are small, arginine-rich, cationic proteins, which can specifically interact with lipid membranes of bacteria, enveloped viruses, or fungi.12 An exception is the recently discovered AMP of the sweat glands, termed dermcidin, which is negatively charged. This processed and secreted protein is active against a broad spectrum of bacteria, including MRSA and rifampin- and isoniazid-resistant Mycobacterium tuberculosis at concentrations of 1 μg/mL.13 Besides cellular systems,5 the most common methods to investigate the activity of AMPs are bulk release assays using © 2014 American Chemical Society

unilamellar vesicles to monitor peptide-induced pore formation in the membrane.14−17 Such experiments, however, can barely be automatized and parallelized. Solid-supported membranes allow for the development of chip-based technologies18,19 and have also been applied to study the interaction with AMPs.20−22 However, dye transfer indicative of peptide-induced pore formation and permeabilization cannot be observed, because a second aqueous compartment is missing. Here, we present a chip-based model membrane assay to be able to visualize the activity of AMPs in a time-resolved manner with the potential to automate the system and monitor pore formation through which water-soluble dyes can pass. As a test case, we used two different AMPs, namely, melittin23,24 and magainin-2,10,25 to evaluate our assay. Melittin is a 26 amino acid long AMP and one of the main components of bee venom. It is discussed in the literature whether melittin penetrates biological membranes according to the toroidal model or according to the carpet model.23 In the carpet model, peptides are orientated parallel to the membrane surface and, at a critical concentration, form toroidal pores but also solubilize the lipid Received: January 26, 2014 Revised: April 7, 2014 Published: April 7, 2014 4767 | Langmuir 2014, 30, 4767−4774



sealing the underlying AAO cavities, as schematically shown in Figure 1B. To be able to spread GUVs, AAO needs to be chemically modified. Without functionalization of the AAO substrate, pore-spanning membranes can hardly be achieved.28 Only the bubble collapse method30 and vesicle spreading under constant flow31 have been reported to facilitate lipid bilayer formation on unmodified AAO surfaces. Here, we used a functionalization strategy based on (3-mercaptopropyl)triethoxysilane that we have recently developed.29 After the entire AAO surface was first silanized, the upper part of the functionalized AAO substrate was protected by evaporating a thin gold layer on top (5−10 nm) and argon and oxygen plasma were applied for 60 s each to remove the silanization from the non-gold-protected AAO pore interior. After the protecting gold layer was removed by an I2/KI solution, oxygen plasma was used to render the (3mercaptopropyl)triethoxysilane-functionalized AAO surface hydrophilic. The functionalized AAO substrate was then first immersed in ethanol and then in phosphate-buffered saline (PBS) to wet the entire cavities to wet the entire cavities. The buffer-immersed and -functionalized AAO was then used to spread sulforhodamine 101 1,2-dihexadecanoyl-sn-glycero-3phosphoethanol-l-amine triethylammonium salt (Texas Red DHPE)-labeled GUVs composed of 1-palmitoyl-2-oleoyl-snglycero-3-phosphocholine (POPC). After the GUVs have spread, the solution was exchanged with fresh buffer solution and the water-soluble fluorescent dye was added. CLSM images of the Texas Red DHPE fluorescence and that of the watersoluble dye fluorescence were taken. The overlay of the fluorescence images (Figure 1C) clearly show the formation of planar lipid bilayer patches on the functionalized AAO substrate as a result of individual GUV rupturing. Because the membranes seal the cavities from the bulk solution, the fluorescence of the water-soluble dye pyranine is excluded from those areas, where the membranes span the buffer-filled cavities and appear black. The hydrophilic functionalization strategy ensures that continuous lipid bilayers are formed lipid bilayers are formed throughout on top of the AAO. From fluorescence recovery after photobleaching experiments, we concluded that the lipids in the top and bottom leaflet of the pore-spanning membranes are fully mobile.29 Because AAO is optically transparent, the water-soluble pyranine dye that is excluded from the underlying cavities of AAO can be monitored simultaneously to the fluorescently labeled membrane by means of three-dimensional CLSM images (z-lines of Figure 1D). Within a typical observation period of about 1.5 h, the pore-spanning membranes prevent the entrance of the watersoluble dye in the underlying cavities. It has been previously shown that pore-spanning membranes on nanometer-sized alumina pores are stable for days and that the stability of such membranes is a function of the pore size.30,32,33 Owing to the long-term stability of the porespanning membranes, they are ideal candidates for the development of an on-chip membrane device to analyze the permeabilizing properties of AMPs. Assay Development: Impact of Magainin-2 on PoreSpanning Membranes. To investigate whether the planar pore-spanning membranes are suited to monitor the effect of AMPs, we set out first experiments with magainin-2. Planar pore-spanning membranes composed of POPC and labeled with Texas Red DHPE were prepared (Figure 2A). After the addition of pyranine into the bulk phase, magainin-2 was added and three-dimensional fluorescence images of Texas Red

membrane in a manner similar to detergents. In the toroidal model, which is also proposed for magainin-2,26 an AMP derived from frog skin, peptides form pores in lipid bilayers at a certain concentration, while the peptides are always in close contact to the lipid head groups and do not solubilize the membrane. To generate a chip-based assay using model membranes in combination with fluorescence readout, we made use of porespanning membranes based on anodic aluminum oxide (AAO). These pore-spanning membranes are stable over long time periods, are partly free-standing, and at the same time, partly immobilized on a solid substrate.27−30 They separate the aqueous phase outside the pores from that inside the pores and, thus, allow for the exclusion of water-soluble dyes. Because AAO is optically transparent, dye transfer across the membrane can be readily monitored, even though not each individual pore can be resolved because its size is well below the optical resolution of a confocal microscope. On the basis of these porespanning membranes, we were able to quantitatively investigate the kinetics of dye transfer across lipid bilayers and to monitor membrane solubilization by means of confocal laser scanning microcopy (CLSM) as a function of the AMP concentration.


Pore-Spanning Lipid Bilayers on AAO. AAO with pore diameters of about 70 nm (Figure 1A) glued on a glass substrate served as the substrate to spread giant unilamellar vesicles (GUVs), resulting in planar pore-spanning membranes

Figure 1. (A) Scanning electron micrograph of AAO showing the cylindrically shaped cavities with a length of several micrometers and a diameter of about 70 nm. (B) Schematic drawing of a pore-spanning lipid bilayer patch (red) sealing the AAO cavities. Only those AAO cavities that are not covered by a pore-spanning membrane are filled with the water-soluble fluorescent dye (green). (C) Overlay of CLSM images of the Texas Red DHPE and pyranine fluorescence. Planar Texas-Red-DHPE-labeled POPC membrane patches (red) are formed on the functionalized AAO substrate as a result of individual GUV rupturing, sealing the underlying cavities of the AAO. Pyranine (green) is only detected in the cavities that are not covered by a pore-spanning membrane. (D) Overlay of fluorescence images showing a z-line profile of a pore-spanning Texas-Red-DHPE-doped POPC bilayer (red) with pyranine (green) in bulk solution. 4768 | Langmuir 2014, 30, 4767−4774



Figure 2. Fluorescence images (top view) of Texas-Red-DHPE-labeled POPC pore-spanning bilayers (A) before and (B) after the addition of 6.6 μM magainin-2. (C) Overlay of fluorescence images showing z-line profiles of pore-spanning Texas-Red-DHPE-labeled POPC bilayers (red) and pyranine (green) at the given time points. The large white box marks the region of the membrane-sealed AAO cavities, from which the time-resolved Texas Red DHPE fluorescence was read out, whereas the small blue and green boxes represent ROIs used for the time-resolved quantitative analysis of the pyranine dye entrance (green, reference ROI, Iref; blue, analyzed ROI, IROI). Upon the addition of 6.6 μM magainin-2 at t = 0 s, pyranine fluorescence (green) becomes visible in the AAO cavities that were sealed by the pore-spanning membrane (red).

cavities, thus giving the maximum fluorescence intensity in the AAO cavities. For each experiment, the time point of peptide addition was set to t = 0. The relative fluorescence intensity as a function of time is calculated using eq 1.

DHPE and pyranine were taken (Figure 2C). Upon the addition of 6.6 μM magainin-2, pyranine starts entering the AAO cavities, indicated by an increase in fluorescence intensity (Figure 2C). This result demonstrates that pores in the lipid bilayers are formed that allow for the entrance of the watersoluble dye into the AAO cavities similar to what has been observed in bulk vesicle release assays.34−36 However, vesicular membranes are highly curved with a significant size distribution of the vesicles, and peptide-induced fusion of vesicles can occur. Moreover, in comparison to bulk vesicle release assays, pore-spanning membranes on a transparent AAO substrate offer the possibility to simultaneously observe the dye entrance into the cavities and the fluorescence of the planar membrane. The observation of the Texas Red fluorescence of the membrane might provide some information about the membrane integrity within the resolution limit of the optical microscope, which cannot be deduced from vesicle assays. The Texas Red DHPE fluorescence (Figure 2B) shows that, even after the pyranine fluorescence has reached its maximum in the AAO cavities, the membrane is still clearly visible. A more quantitative analysis has been performed, as detailed below. Kinetics Analysis of Magainin-2 Permeabilization. From the time-dependent change in fluorescence intensity of pyranine, the kinetics of dye entrance into the cavities can be read out. The time-dependent change in fluorescence intensity [IROI(t)] in a chosen region of interest (ROI), marked by the blue box in Figure 2C, was normalized by reading out the fluorescence intensity at t = 0 (I0) and standardized against the fluorescence intensity of a reference ROI [Iref(t)], marked by the green box in Figure 2C, at every given time point. The reference ROI was chosen in a membrane-free area on the AAO substrate, where the fluorescent dye can readily diffuse into the

Irel(t ) =

IROI(t ) − I0 Iref (t ) − I0


To evaluate the readout procedure and the minimum possible time resolution of the assay, which was limited by the z-line scans providing a good signal-to-noise ratio, we added 0.5 mM Triton X-100 to a pore-spanning membrane and monitored Irel as a function of time (see Figure S1 of the Supporting Information). Triton X-100 is known to permeabilize membranes quickly, resulting in pores through which the water-soluble dye can pass. Irel increased from 0 to 1 within 50 s, indicating that all cavities are filled with pyranine after 50 s. Figure 3A shows the kinetics of pyranine entrance into the cavities [Irel(t)] for five different magainin-2 concentrations ranging from 0.7 to 6.6 μM. While there is almost no dye entrance at a concentration of 0.7 μM, increasing concentrations result in an increased and faster dye entrance. It becomes obvious that a threshold concentration is required to achieve significant membrane permeabilization, which is between 0.7 and 1.4 μM. This result is in line with the results obtained by Almeida et al.37 Only if a certain threshold concentration of magainin-2 in solution is exceeded, pores in the membrane are formed. To quantitatively evaluate the observed sigmoidal kinetics, we fitted a Boltzmann function (eq 2) to the data (Figure 3A) Irel(t ) =

I0 − Imax


1 + exp 4769

t − t1/2 τ


+ Imax (2) | Langmuir 2014, 30, 4767−4774



One advantage of the presented setup is that, besides the possibility to observe the kinetics of dye entrance into the cavities as a result of membrane pore formation, the integrity of the membrane itself can be observed in a time-dependent manner by monitoring the fluorescence intensity Irel,mem of the Texas Red DHPE membrane fluorophore (Figure 3E). The results demonstrate that, at magainin-2 concentrations of 2.8 and 5.6 μM, where a significant dye entrance is observed, the Texas Red DHPE membrane fluorescence remains rather constant over time, except for some photobleaching, indicating that lipid material is not removed from the surface to a large extent. Only at the highest used magainin-2 concentration of 6.6 μM (see also Figure 2B) is a decrease in membrane fluorescence observed, indicating the loss of lipid material from the surface. The kinetics together with the stability of the membrane attached to the support can be discussed in light of the proposed toroidal model described for magainin-2.26 In the toroidal model, it is proposed that the kinetics of pore formation in membranes is governed by first binding of peptides in a parallel orientation to the membrane interface until a critical threshold concentration is reached, after which they self-organize in the membrane to form a permeation pathway. Above this threshold concentration, a transmembrane pore appears associated with the peptides, adopting a more perpendicular orientation. Applying this model to our experimental findings, where we observe pore formation by reading out pyranine dye entrance into the AAO cavities, we would estimate such a threshold concentration to be >0.7 μM. The toroidal model further proposes that a larger magainin-2 concentration in solution results in more membrane-bound peptides with a perpendicular orientation, leading to more and larger toroidal pores, respectively. The time point at which these toroidal pores are formed as a result of the selforganization of the peptides in the membrane is influenced by the magainin-2 concentration. In our assay, this concentrationdependent pore formation is reflected by a decreasing t1/2 with an increasing magainin-2 concentration. The observation that the velocity of dye entrance reflecting the number and size of the pores is also influenced by the magainin-2 concentration, suggests that a larger peptide concentration leads to the formation of more and/or larger pores, which is reasonable in light of more membrane-bound peptides at higher concentrations. If the once formed pores were capable of laterally diffusing within the membrane patch, in principle, all cavities were expected to be filled with the fluorescent dye. However, we observed that, at lower peptide concentrations, the membrane is not fully permeabilized, suggesting that the pores are not fully mobile. The assumption of toroidal pore formation in the case of magainin-2 might also be reflected in the observation that the integrity of the membrane remains largely intact even at higher peptide concentrations. We cannot rule out that morphological changes in the membrane occur, which cannot be resolved by the membrane fluorescence. However, we can safely state that, up to a magainin-2 concentration of 5.6 μM, no significant loss of lipid material into the bulk solution is observed, which would be expected if magainin-2 significantly solubilized the membrane. To gather further information about the pore sizes that are formed by magainin-2, we used fluorescently labeled 70 kDa dextran molecules (Dex70) instead of pyranine (Pyr) to monitor their entrance into the AAO pores upon magainin-2

Figure 3. (A) Time courses of Irel as a function of different magainin-2 concentrations in solution. The black solid line is the result of the fit of eq 2 to the data. A pyranine concentration of 1 mM was added to the bulk phase. (B) Imax (▲), (C) center point t1/2 (●), and (D) τ (⧫) as a function of the magainin-2 concentration in solution. For each data point, at least two independent membrane preparations were used. The error bars show the standard error of the mean. (E) Time courses of the Texas Red DHPE fluorescence intensity Irel,mem as a function of different magainin-2 concentrations in solution.

which allowed us to extract the maximum intensity Imax as well as the center time t1/2 together with the slope τ of the tangent at t1/2. The Imax values gradually increase with increasing magainin-2 concentrations from Imax = 0.7 at 1.4 μM to Imax = 1 at concentrations larger than 5 μM (Figure 3B). This result suggests that the membranes become gradually permeabilized as a function of the magainin-2 concentration, with some porespanning membranes remaining stable without pore formation. Also, the center time t1/2 exhibits a dependence upon the magainin-2 concentration (Figure 3C). An increase in the magainin-2 concentration decreases t1/2, reflecting that the lag time between magainin-2 addition and maximum dye entrance, which indicates permeabilization of the membrane, is decreased with an increasing magainin-2 concentration. In addition to the dependence of the lag phase, also the velocity of the dye entrance into the cavities at t1/2 is altered by the magainin-2 concentration, being faster at larger magainin-2 concentrations (Figure 3D). 4770 | Langmuir 2014, 30, 4767−4774



addition. The minimum possible time resolution of the setup using this large dye molecule was again analyzed by permeabilizing the membrane with Triton X-100 (see Figure S1 of the Supporting Information), indicating that the observed permeabilization kinetics observed in the presence of magainin2 are not limited by the diffusion of the dye into the cavities or the resolution of the z-scan imaging. Figure 4 shows that, at a magainin-2 concentration of 4.2 μM, the center time t1/2 is shifted from 1330 s in the presence

Figure 4. Time courses of Irel in the presence of pyranine (Pyr) and FITC−dextran (70 kDa, Dex70) as a function of the magainin-2 concentration.

of pyranine to t1/2 = 2600 s in the presence of fluorescein isothiocyanate (FITC)−dextran (70 kDa). This finding suggests that the pores formed by the assembly of magainin-2 grow with time, so that, only after a significant lag time, when the peptides arranged into larger pores, the fluorescently labeled dextran (Stokes−Einstein radius of 6.4 nm)38 is capable of entering the AAO cavities. If a lower magainin-2 concentration of 2.4 μM is applied, which is a sufficiently high concentration to produce small pores to let pyranine pass, the magainin-2-induced pores do not grow to a size to allow for FITC−dextran (70 kDa) to enter the AAO cavities. Kinetics Analysis of Melittin Permeabilization. To prove whether the developed assay is suited for different AMPs, we performed the same experiments using melittin, instead of magainin-2 (Figure 5A). From the concentration-dependent kinetics of pyranine dye entrance into the AAO cavities, it becomes obvious that the minimum concentration required for a significant permeabilization of the membrane is by about a factor of 4 smaller for melittin than for magainin-2, which is in line with concentration ranges reported in the literature using vesicles.16,21,39,40 While, at 0.13 μM melittin, only a small dye entrance is observed, a concentration of 0.35 μM is already sufficient to observe the maximum dye entrance (Figure 5B). The center time t1/2 and the slope τ at t1/2 decrease with increasing melittin concentrations (panels C and D of Figure 5). Of note, the center time, which is a measure of the lag phase between melittin addition and its permeabilizing action on the membrane, is much smaller compared to that found for magainin-2. At a concentration of 1.4 μM, t1/2 reads 2160 s for magainin-2, while it is by a factor of 30 smaller (t1/2 = 75 s) for melittin. Also, the velocity of dye entrance, i.e., τ, is by a factor of about 55 larger for melittin compared to magainin-2 at the same concentration of 1.4 μM. This observation might be attributed to a larger number of pores, which are expected to increase with larger melittin concentrations.16,17,24 The impact of melittin on membranes is discussed in terms of a carpet-like mechanism, leading to micellization, particularly

Figure 5. (A) Time courses of Irel as a function of different melittin concentrations in solution. A pyranine concentration of 1 mM was added to the bulk phase. (B) Imax (▲), (C) center point t1/2 (●), and (D) τ (⧫) as a function of the melittin concentration in solution obtained by fitting eq 2 to the data. For each data point, at least two independent membrane preparations were used. The error bars show the standard error of the mean. (E) Time courses of the Texas Red DHPE fluorescence intensity Irel,mem as a function of different melittin concentrations in solution.

at higher melittin concentrations.24 The model proposes that a certain minimum concentration is required to permeabilize the membrane by melittin. At low concentrations, the peptide is preferentially aligned parallel to the membrane surface. If a certain peptide surface concentration is reached, the peptides insert into the lipid bilayer perpendicular to the membrane plane,15,23,41 leading to permeating pores. Besides these permeation pathways, the effect of melittin on membranes has been reported to also include removal of lipid material from the surface.22 In light of the proposed model for melittin, one could interpret our observations as follows: because the center points 4771 | Langmuir 2014, 30, 4767−4774



AAO consists of two anodization steps. The polished Al foils were first anodized for 2.5 h at 40 V and 1.5 °C. Then, the first less-ordered AAO layer was removed by incubation in 5% H3PO4 for 2−3 h, and the remaining pre-structured Al surface was anodized for a second time under the same conditions as described above. Finally, the remaining Al was dissolved in an acidic copper solution (17 g/L CuCl2 in 1:1 H2O/HCl). For silanization, AAO was glued on a glass slide with an ultraviolet (UV) curable adhesive (Norland Optical Adhesive 83H) dissolved in tetrahydrofuran (1:10). The pore diameter was widened for 50 min in 5% H3PO4 to achieve a radius of about 70 nm. For the orthogonal surface functionalization, the mounted AAO was treated with oxygen plasma (plasma cleaner PDC 32 G-2, Harrick, Ithaka, NY) for 1 min to maximize the amount of hydroxyl groups. The glass slides with AAO were then silanized by incubating them in a sealed chamber with 50 μL of (3-mercaptopropyl)triethoxysilane at 135 °C at a constant vacuum for at least 3 h. To protect the surface functionalization on the pore rims, a thin gold layer (10 nm) was evaporated (coating system MED020, Leica, Wetzlar, Germany). Preparation of GUVs. GUVs were prepared according to the procedure by Angelova et al.44 Briefly, 25 μL of lipid solution (2 mg/ mL) in trichloromethane was added to each of two ITO-coated glass slides and dried under vacuum for at least 3 h. Texas Red DHPE (0.5 mol %) was used for fluorescently labeling the membranes. The ITO slides were connected via copper tape on each side, and a chamber was formed with a silicon spacer. The lipid film was rehydrated in 0.3 M sucrose solution, and an alternating current (AC) field was applied for 3 h at 3 V and 5 Hz. Preparation of Pore-Spanning Lipid Bilayers on AAO Substrates. Prior to preparation of pore-spanning lipid bilayers, the surface functionalization of the pore interior was removed by treating AAO with Ar and O2 plasma (1 min). The thin Au layer was removed by immersing AAO in I2/KI (0.5% KI and 0.125% I2 in H2O), followed by rinsing with water and ethanol. The dried sample was then O2 plasma treated (1 min) to oxidize the silane to silanol and rendered the surface of the pore rims hydrophilic. AAO was immediately immersed in ethanol, which was then replaced by PBS. A total of 50− 100 μL of GUV suspension was immediately added and incubated for at least 15 min. After preparation of the pore-spanning membranes, AAO was first rinsed with buffer, then 1 mM pyranine or 5 μM FITC− dextran (70 kDa) was added to the buffer, and three-dimensional CLSM images (z-scans) were taken. Then, AMP (melittin or magainin-2) dissolved in buffer was added to the solution (total volume of 3 mL) to obtain the given concentration while stirring. CLSM. Three-dimensional fluorescence images (z-stacks) of the pore-spanning lipid bilayers and the water-soluble dyes were taken with a confocal laser scanning microscope (LSM 710, Carl Zeiss, Jena, Germany) equipped with a water immersion objective W PlanApochromat 63×/1.0 na (Carl Zeiss, Jena, Germany). Texas Red DHPE was excited at λex = 594 nm, and the emission was detected at λem = 605−690 nm. Pyranine and FITC−dextran were excited at λex = 488 nm, and the emission was detected at λem = 490−580 nm.

are much smaller and the velocities of dye entrance at t1/2 are much faster compared to those found for magainin-2, it is assumed that, in the case of melittin, more and larger permeabilizing pores are formed on a faster time scale. These permeabilizing pores can, in part, be a result of a melittininduced micellization of the membrane, which would result in a partial loss of lipid material into the bulk solution. We indeed observed significant and time-dependent removal of lipid material from the AAO surface as a function of the melittin concentration in solution (Figure 5E and see Figure S2 of the Supporting Information). Several studies have been attempted to monitor the structure and function of melittin-induced pores, and their results show that melittin forms pores that have a rather wide distribution of sizes, so that one cannot distinguish between more and/or larger pores.24 We tried to estimate the formed pore sizes by performing the dye translocation experiments in the presence of FITC−dextran (70 kDa). It turned out that, in contrast to what has been observed for magainin-2, the kinetics of FITC−dextran (70 kDa) entrance were very similar to those found for pyranine at the same melittin concentration (see Figure S3 of the Supporting Information). This result demonstrates that the melittin-induced permeating pores because of micellization are large enough for FITC−dextran (70 kDa) to pass, even at lower melittin concentrations, consistent with the finding that, already at lower melittin concentrations, part of the membranes becomes removed from the AAO surface (Figure 5E). In summary, the results demonstrate that our assay provides a tool to investigate the concentration range of AMP activity and delivers some information on how the AMP acts on the membrane. Indeed, the concentration ranges in which melittin and magainin-2 permeabilize the pore-spanning membranes are in close agreement with recently reported ED50 values found for these two peptides using cell viability assays and K+ efflux of S. aureus cells.42

CONCLUSION We developed a chip-based membrane assay to investigate the permeabilizing properties of AMPs. Planar, non-curved membranes spanning the pores of transparent AAO allow for the monitoring of simultaneous changes in membrane integrity and the transfer of a water-soluble dye indicative of the formation of permeabilization pathways through the membrane. This setup in conjunction with optical microscopy has the potential to screen, identify, and analyze AMPs to gather information about their membrane permeabilizing properties. Because there is a growing demand for the development of new antimicrobial agents replacing standard antibiotics, assays to test the activity of these molecules are greatly needed.


S Supporting Information *


Irel(t) before and after the addition of Triton X-100 (Figure S1), fluorescence images of Texas-Red-DHPE-labeled POPC pore-spanning bilayers before and after the addition of melittin (Figure S2), and Irel(t) in the presence of pyranine and FITC− dextran (70 kDa) as a function of the melittin concentration (Figure S3). This material is available free of charge via the Internet at

Materials. Sulforhodamine 101 1,2-dihexadecanoyl-sn-glycero-3phosphoethanol-l-amine triethylammonium salt (Texas Red DHPE), melittin, magainin-2, and FITC−dextran (70 kDa) were purchased from Sigma-Aldrich (Taufkirchen, Germany). POPC was from Avanti Polar Lipids (Alabaster, AL). Indium tin oxide (ITO) slides were from Präzisions Glas and Optik GmbH (Iserlohn, Germany), and (3mercaptopropyl)triethoxysilane was from ABCR (Karlsruhe, Germany). Pyranine was purchased from ACROS Organics (Geel, Belgium). Sucrose was from Carl Roth GmbH (Karlsruhe, Germany). AAO Substrates. AAO was prepared as previously described.43 Briefly, after annealing the Al foils at 500 °C overnight, they were electrochemically polished at 65 °C and 25 V in concentrated H2SO4/ 85% H3PO4/H2O (1:1:1) twice for 15 min. The fabrication of the


Corresponding Author

*E-mail: [email protected]. 4772 | Langmuir 2014, 30, 4767−4774



Author Contributions

(21) Shaw, J. E.; Epand, R. F.; Hsu, J. C. Y.; Mo, G. C. H.; Epand, R. M.; Yip, C. M. Cationic peptide-induced remodelling of model membranes: Direct visualization by in situ atomic force microscopy. J. Struct. Biol. 2008, 162, 121−138. (22) Steinem, C.; Galla, H.-J.; Janshoff, A. Interaction of melittin with solid supported membranes. Phys. Chem. Chem. Phys. 2000, 2, 4580− 4585. (23) 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. (24) Raghuraman, H.; Chattopadhyay, A. Melittin: A membraneactive peptide with diverse functions. Biosci. Rep. 2007, 27, 189−223. (25) Zasloff, M. Magainins, a class of antimicrobial peptides from Xenopus skin: Isolation, characterization of two active forms, and partial cDNA sequence of a precursor. Proc. Natl. Acad. Sci. U. S. A. 1987, 84, 5449−5453. (26) Ludtke, S. J.; He, K.; Heller, W. T.; Harroun, T. A.; Yang, L.; Huang, H. W. Membrane pores induced by magainin. Biochemistry 1996, 35, 13723−13728. (27) Lazzara, T. D.; Mey, I.; Steinem, C.; Janshoff, A. Benefits and limitations of porous substrates as biosensors for protein adsorption. Anal. Chem. 2011, 83, 5624−5630. (28) Lazzara, T. D.; Kliesch, T.-T.; Janshoff, A.; Steinem, C. Orthogonal functionalization of nanoporous substrates: Control of 3D surface functionality. ACS Appl. Mater. Interface 2011, 3, 1068−1076. (29) Lazzara, T. D.; Carnarius, C.; Kocun, M.; Janshoff, A.; Steinem, C. Separating attoliter-sized compartments using fluid pore-spanning lipid bilayers. ACS Nano 2011, 5, 6935−6944. (30) Mager, M. D.; Melosh, N. A. Nanopore-spanning lipid bilayers for controlled chemical release. Adv. Mater. 2008, 20, 4423−4427. (31) Venkatesan, B. M.; Polans, J.; Comer, J.; Sridhar, S.; Wendell, D.; Aksimentiev, A.; Bashir, R. Lipid bilayer coated Al2O3 nanopore sensors: Towards a hybrid biological solid-state nanopore. Biomed. Microdevices 2011, 13, 671−682. (32) Römer, W.; Lam, Y. H.; Fischer, D.; Watts, A.; Fischer, W. B.; Göring, P.; Wehrspohn, R. B.; Gösele, U.; Steinem, C. Channel activity of a viral transmembrane peptide in micro-BLMs: Vpu1−32 from HIV1. J. Am. Chem. Soc. 2004, 126, 16267−16274. (33) Weiskopf, D.; Schmitt, E. K.; Klühr, M. H.; Dertinger, S. K.; Steinem, C. Micro-BLMs on highly ordered porous silicon substrates: Rupture process and lateral mobility. Langmuir 2007, 23, 9134−9139. (34) Matsuzaki, K.; Harada, M.; Handa, T.; Funakoshi, S.; Fujii, N.; Yajima, H.; Miyajima, K. Magainin 1-induced leakage of entrapped calcein out of negatively-charged lipid vesicles. Biochim. Biophys. Acta 1989, 981, 130−134. (35) Grant, E.; Beeler, T. J.; Taylor, K. M. P.; Gable, K.; Roseman, M. A. Mechanism of magainin 2a induced permeabilization of phospholipid vesicles. Biochemistry 1992, 31, 9912−9918. (36) Gregory, S. M.; Pokorny, A.; Almeida, P. F. Magainin 2 revisited: A test of the quantitative model for the all-or-none permeabilization of phospholipid vesicles. Biophys. J. 2009, 96, 116− 131. (37) Almeida, P. F.; Pokorny, A. Mechanisms of antimicrobial, cytolytic, and cell-penetrating peptides: From kinetics to thermodynamics. Biochemistry 2009, 48, 8083−8093. (38) Goins, A. B.; Sanabria, H.; Waxham, M. N. Macromolecular crowding and size effects on probe microviscosity. Biophys. J. 2008, 95, 5362−5373. (39) Arnusch, C. J.; Branderhorst, H.; de Kruijff, B.; Liskamp, R. M. J.; Breukink, E.; Pieters, R. J. Enhanced membrane pore formation by multimeric/oligomeric antimicrobial peptides. Biochemistry 2007, 46, 13437−13442. (40) Mally, M.; Majhenc, J.; Svetina, S.; Zeks, B. The response of giant phospholipid vesicles to pore-forming peptide melittin. Biochim. Biophys. Acta 2007, 1768, 1179−1189. (41) Lee, M.-T.; Chen, F.-Y.; Huang, H. W. Energetics of pore formation induced by membrane active peptides. Biochemistry 2004, 43, 3590−3599.

Henrik Neubacher and Ingo Mey contributed equally to this work. Notes

The authors declare no competing financial interest.

ACKNOWLEDGMENTS Claudia Steinem gratefully acknowledges financial support from the Deutsche Forschungsgemeinschaft (DFG, SFB 803).


(1) Bayley, H.; Cremer, P. S. Stochastic sensors inspired by biology. Nature 2001, 413, 226−230. (2) Yamazaki, V.; Sirenko, O.; Schafer, R. J.; Nguyen, L.; Gutsmann, T.; Brade, L.; Groves, J. T. Cell membrane array fabrication and assay technology. BMC Biotechnol. 2005, 5, 18. (3) Richard, J. P.; Melikov, K.; Vives, E.; Ramos, C.; Verbeure, B.; Gait, M. J.; Chernomordik, L. V.; Lebleu, B. Cell-penetrating peptides. A reevaluation of the mechanism of cellular uptake. J. Biol. Chem. 2003, 278, 585−590. (4) Joliot, A.; Prochiantz, A. Transduction peptides: From technology to physiology. Nat. Cell Biol. 2004, 6, 189−196. (5) Brogden, K. A. Antimicrobial peptides: Pore formers or metabolic inhibitors in bacteria? Nat. Rev. Microbiol. 2005, 3, 238−250. (6) Henriques, S. T.; Melo, M. N.; Castanho, M. A. R. B. Cellpenetrating peptides and antimicrobial peptides: How different are they? Biochem. J. 2006, 399, 1−7. (7) Yeung, A. T. Y.; Gellatly, S. L.; Hancock, R. E. W. Multifunctional cationic host defence peptides and their clinical applications. Cell. Mol. Life Sci. 2011, 68, 2161−2176. (8) Rice, L. B. Unmet medical needs in antibacterial therapy. Biochem. Pharmacol. 2006, 71, 991−995. (9) Schittek, B.; Hipfel, R.; Sauer, B.; Bauer, J.; Kalbacher, H.; Stevanovic, S.; Schirle, M.; Schroeder, K.; Blin, N.; Meier, F.; Rassner, G.; Garbe, C. Dermcidin: A novel human antibiotic peptide secreted by sweat glands. Nat. Immunol. 2001, 2, 1133−1137. (10) Zasloff, M. Antimicrobial peptides of multicellular organisms. Nature 2002, 415, 389−395. (11) Hancock, R. E.; Scott, M. G. The role of antimicrobial peptides in animal defenses. Proc. Natl. Acad. Sci. U. S. A. 2000, 97, 8856−8861. (12) Hristova, K.; Selsted, M. E.; White, S. H. Critical role of lipid composition in membrane permeabilization by rabbit neutrophil defensins. J. Biol. Chem. 1997, 272, 24224−24233. (13) Paulmann, M.; Arnold, T.; Linke, D.; Ö zdirekcan, S.; Kopp, A.; Gutsmann, T.; Kalbacher, H.; Wanke, I.; Schuenemann, V. J.; Habeck, M.; Bürck, J.; Ulrich, A. S.; Schittek, B. Structure−activity analysis of the dermcidin-derived peptide DCD-1L, an anionic antimicrobial peptide present in human sweat. J. Biol. Chem. 2012, 287, 8434−8443. (14) van den Bogaart, G.; Mika, J. T.; Krasnikov, V.; Poolman, B. The lipid dependence of melittin action investigated by dual-color fluorescence burst analysis. Biophys. J. 2007, 93, 154−163. (15) van den Bogaart, G.; Guzmán, J. V.; Mika, J. T.; Poolman, B. On the mechanism of pore formation by melittin. J. Biol. Chem. 2008, 283, 33854−33857. (16) Kokot, G.; Mally, M.; Svetina, S. The dynamics of melittininduced membrane permeability. Eur. Biophys. J. 2012, 41, 461−474. (17) Matsuzaki, K.; Yoneyama, S.; Miyajima, K. Pore formation and translocation of melittin. Biophys. J. 1997, 73, 831−838. (18) Castellana, C. P. S.; Edward, T. Solid supported lipid bilayers: From biophysical studies to sensor design. Surf. Sci. Rep. 2006, 61, 429−444. (19) Reimhult, E.; Kumar, K. Membrane biosensor platforms using nano- and microporous supports. Trends Biotechnol. 2008, 26, 82−89. (20) Machán, 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. 4773 | Langmuir 2014, 30, 4767−4774



(42) Nakao, S.; Komagoe, K.; Inoue, T.; Katsu, T. Comparative study of the membrane-permeabilizing activities of mastoparans and related histamine-releasing agents in bacteria, erythrocytes, and mast cells. Biochim. Biophys. Acta 2011, 1808, 490−497. (43) Lazzara, T. D.; Aaron Lau, K. H.; Knoll, W. Mounted nanoporous anodic alumina thin films as planar optical waveguides. J. Nanosci. Nanotechnol. 2010, 10, 4293−4299. (44) Angelova, M. I.; Soléau, S.; Méléard, P.; Faucon, F.; Bothorel, P. In Progress in Colloid and Polymer Science; Helm, C., Lösche, M., Möhwald, H., Eds.; Steinkopff and Springer-Verlag: Darmstadt, Netherlands, 1992; Vol. 89; pp 127−131.

4774 | Langmuir 2014, 30, 4767−4774

Permeabilization assay for antimicrobial peptides based on pore-spanning lipid membranes on nanoporous alumina.

Screening tools to study antimicrobial peptides (AMPs) with the aim to optimize therapeutic delivery vectors require automated and parallelized sampli...
2MB Sizes 1 Downloads 3 Views