Characterizing α-Helical Peptide Aggregation on Supported Lipid

Aug 27, 2014 - Joshua A. Jackman , Eric Linardy , Daehan Yoo , Jeongeun Seo , Wei Beng Ng , Daniel J. Klemme , Nathan J. Wittenberg , Sang-Hyun Oh ...
0 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/ac

Characterizing α‑Helical Peptide Aggregation on Supported Lipid Membranes Using Microcantilevers Jinghui Wang,† Kai-Wei Liu,† and Sibani Lisa Biswal* Department of Chemical and Biomolecular Engineering, Rice University, Houston, Texas 77005, United States S Supporting Information *

ABSTRACT: We report the use of lipid membrane-coated microcantilevers to probe the interactions between phospholipid membranes and membrane-active peptides. This sensing method integrates two well-developed techniques: solid-supported lipid bilayers (SLBs) and microcantilever sensors. SLBs are prepared on the silicon dioxide surface of the microcantilevers using a vesicle fusion method. As molecules adsorb onto the surface of the microcantilever, the microcantilever bends due to the induced compressive or tensile stresses, which result from the surface free energy change. Real-time surface stress changes in the SLB due to interactions with small molecules can be detected by monitoring the deflection of the microcantilever. We investigate the mechanism for the interaction between SLBs and PEP1, a synthetic amphipathic peptide resembling a segment of the nonstructural protein (NS5A) of the hepatitis C virus. Initially, the PEP1 peptides adsorb onto the lipid membranes, and then at a critical concentration, the peptides begin to aggregate and form pores; finally, the peptides destabilize and induce solubilization of the supported membranes. The membrane-coated microcantilever sensor is capable of characterizing the kinetics and dynamics of this process with great sensitivity.

U

leaflet. Cho’s AFM images showed that the sequential introduction of AH to the SLBs caused membrane thinning at AH concentrations ranging from 1.63 μM to 3.25 μM,6 which indicates that the membrane laterally expands according to Huang’s geometric model.9 Thus, the virus-mimetic attack of NS5A-derived AH for rupturing lipid membranes provides insights into binding mechanisms. Furthermore, this attack provides a novel approach for patterning SLBs onto surfaces that do not easily rupture vesicles, such as gold and TiO2 surfaces; difficulties have been reported in spontaneously forming SLBs on these surfaces without external forces.8,10 Instead of purifying this AH segment of the NS5A protein from HCVs, we use a synthetic analogue, PEP1 (synthesized by AnaSpec, San Jose, CA). Similar to the natural NS5A-derived AH, PEP1 exhibits the typical characteristics of amphipathic molecules: it partially penetrates the lipid bilayer and interacts with both polar hydrophilic lipid headgroups and hydrophobic tails. In addition, the PEP1 peptide displays potent antiviral activity against certain viruses.11 PEP1 has been found to destabilize viral membranes and lyse virions.2 Several binding and disruption processes have been proposed empirically; however, a physical understanding is still lacking.6 Unlike PEP1, other well-characterized amphipathic helical peptides, such as melittin and magainin, belong to the family of antibacterial

nderstanding peptide−membrane interactions is important for elucidating biomolecular mechanisms, such as those underlying membrane fusion, cell signaling, and the therapeutic delivery of antibiotic and antiviral drugs. Membrane-active peptides are proteins that are known for their association with lipid membranes. These peptides interact predominantly with the hydrocarbon region of lipid membranes, the polar headgroup region, or both regions of the bilayer, depending on the peptide’s hydropathicity. An interesting amphipathic helix (AH) peptide is the AH segment derived from the N-terminal end of nonstructural protein NS5A of the hepatitis C virus. The hepatitis C virus (HCV) infects more than 170 million people worldwide and places those who are exposed at risk for chronic liver disease;1,2 therefore, understanding the infectious mechanism of HCV is vital. The association of the NS5A-derived AH peptide with lipid membranes is an essential step in HCV viral infection. The AH viral attack is able to induce lysis of lipid membranes. Blocking the membrane-binding pathway of AH is a promising therapeutic strategy.3 Cho et al. and Chah et al. investigated the binding mechanism of the interaction of this AH peptide with model lipid membranes, either in lipid vesicles (SUVs or LUVs) or supported lipid membranes, with analytical tools, such as quartz crystal microbalance with dissipation (QCM-D), dynamic light scattering (DLS), reflectometry, and atomic force microscopy (AFM).4−8 The NS5A-derived AH peptide is thought to bind parallel to the lipid membrane surface and penetrate the lipid membrane, leading to expansion of the outer © 2014 American Chemical Society

Received: April 14, 2014 Accepted: August 13, 2014 Published: August 27, 2014 10084

dx.doi.org/10.1021/ac501343b | Anal. Chem. 2014, 86, 10084−10090

Analytical Chemistry

■

peptides.9,12−15 These peptides are known for adsorbing onto lipid membranes and reorienting to form transmembrane pores, leading to higher membrane permeability or even membrane disruption. We refer to the theories and models developed for these amphiphilic peptides to explain the membrane binding and possible pore formation of PEP1. More specifically, we are interested in measuring the PEP1-induced lateral expansion of the membrane, because this expansion is an indicator of the insertion of foreign molecules into the membrane. Here, we describe how the lateral expansion of the membrane in response to peptide adsorption and pore formation can be detected using microcantilevers. Microcantilever-based sensors are increasingly being used to probe nanoscale quantities of mass, heat, and surface stress for biomolecular interactions.16 In dynamic mode, Braun et al. reported measuring the mass of virus binding to membrane proteins using proteoliposomes immobilized on microcantilevers.17 In stress-based cantilever applications, one surface of the cantilever beam is rendered sensitive to a specific target molecule of interest, while the opposing surface is chemically passivated. When these target molecules interact with the sensitized surface of the cantilever, a surface stress can be induced. The difference in surface stress induced on the sensitive relative to the passive surface of the cantilever results in a measurable mechanical deflection. Previously, microcantilevers have been shown to be useful for sensing peptide binding and reconstruction on surfaces; for example, amyloid fibrils of insulin grow in the in-plane direction and generate a tensile surface stress of 20 mN/m within 2 h on microcantilever surfaces.18,19 Ghatkesar et al. observed the binding of melittin molecules to lipid vesicle membranes deposited on the gold surfaces of microcantilevers.20 These findings suggest possible techniques for monitoring protein aggregation using microcantilevers; however, there has not been a quantitative explanation of how the surface stress can be related to changes in the surface-bound proteins. This study establishes a method for quantifying the peptide-induced mechanical response of microcantilevers. Previously, we provided a systematic characterization of the adsorption of SLBs onto a microcantilever surface and the analysis of the electrostatic and hydrophobic contributions that participate in the changes in adsorption free energy.21 In addition, the insertion and solubilization effects of amphiphilic surfactants on SLBs were detected using lipid membranecoated microcantilevers.22 More recently, the solubility of amphiphilic block copolymers with lipid membranes have been shown to be directly correlated with cantilever deflection.23 In the present paper, we describe a practical application of our previously developed lipid membrane-coated cantilever sensors for biomolecular recognition at model membrane interfaces. This study focuses on the nonspecific α-helix-induced peptide interaction with a supported lipid bilayer. A lipid membranecoated microcantilever has the potential to probe membrane motions in lateral directions, thus providing information other than the thickness and bound mass, which are studied with ellipsometry and quartz crystal microbalance, respectively. Here, we will unravel the concentration-dependent interactions between amphipathic PEP1 peptides and phospholipid membranes. The response of the lipid membrane-coated cantilevers is analyzed based on the free energy change of this interaction between model membranes and peptides.

Article

MATERIALS AND METHODS

Lipid Vesicle Preparation. A zwitterionic lipid, 1palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC), was used to form a neutral charged lipid bilayer. POPC was chosen since it has been previously shown to capture the basic structure and dynamics of plasma membranes.24 The lipids were purchased from Avanti Polar Lipids (Alabaster, AL). Vesicles were prepared by the extrusion method.25 Briefly, lipids were dissolved at a concentration of 5 mg/mL in chloroform. The chloroform was evaporated under a nitrogen stream. The resulting lipid film was then dried in a vacuum chamber for 2 h and then hydrated in a solution composed of 0.5:99.5 (vol %) of DMSO and 0.25 mL pH 7.4 phosphate buffered saline (PBS, Sigma−Aldrich) solutions prepared with ultrapure deionized water (Barnstead Nanopure system, Thermo Fisher Scientific), followed by vortexing. The solution was then extruded 40 times through a 100-nm polycarbonate membrane using a mini-extruder (Avanti Polar Lipids), resulting in a translucent solution of large unilamellar vesicles (LUVs) that were ∼100 nm in size. The vesicle solution was further diluted with 9 parts of PBS to 1 part of the freshly extruded vesicle solution and stored at 4 °C until use. Note that the final vesicle concentration may have been lower than the initially desired concentration, because of lipid loss to the filter membranes after extrusion; however, the concentration was well above the threshold needed to achieve full surface coverage of the SLBs. Amphipathic Helix Peptide Preparation. The peptide, PEP1, was synthesized by AnaSpec (San Jose, CA). PEP1 has 31 amino acids (MW = 3804.3 g/mol) in its sequence: H− Ser−Gly−Ser−Trp−Leu−Arg−Asp−Val−Trp−Asp−Trp− Ile−Cys−Thr−Val−Leu−Thr−Asp−Phe−Lys−Thr−Trp− Leu−Gln−Ser−Lys−Leu−Asp−Tyr− Lys−Asp−NH2. PEP1 was used as delivered without further purification. The peptide was dissolved in a small amount of dimethyl sulfoxide (DMSO, (CH3)2SO, MW = 78.14 g/mol, EMD Chemicals, USA) and then slowly diluted in phosphate buffered saline (PBS, Sigma− Aldrich, USA). The final peptide solution was prepared in an eluent mixture of DMSO and PBS (0.5:99.5 (vol %)). The peptide solution was kept refrigerated at 4 °C for no more than 3 days prior to the experiments. PEP1 is an amphipathic helix peptide that forms a helix structure when adsorbed onto lipid membranes. Figure S1 (shown in Supporting Information) shows a helical wheel diagram26,27 that illustrates the hydrophilic and hydrophobic surfaces on PEP1. Preparation of Microcantilever Surface. Cantilever chips were purchased from Concentris GmbH (Basel, Switzerland). The cantilevers were 500 μm long, 100 μm wide, and 1 μm thick. Each chip contained eight rectangular silicon cantilevers, each with a spring constant of 0.026 N/m. The cantilevers were coated with 3 nm of titanium, followed by a 20 nm layer of gold, resulting in a bimetallic structure. The cantilever arrays were placed in an ultraviolet (UV) ozone cleaner for 5 min under 5 psi oxygen to clean the front of the gold surface and generate a hydrophilic silicon dioxide surface on the back surface. Each cantilever was individually functionalized using glass microcapillaries. The surface functionalization process typically lasted 2 h. To prevent preferential binding to the gold side of the cantilever beam, polyethylene glycol (PEG) polymers were used. A dithiolaromatic−PEG molecule (C25H44O6S2, MW = 504.74 g/mol, Sensopath Technologies, Bozeman, MT) was used to prevent vesicle adsorption to the 10085

dx.doi.org/10.1021/ac501343b | Anal. Chem. 2014, 86, 10084−10090

Analytical Chemistry

■

Article

RESULTS Lipid Deposition onto Functionalized Microcantilever Surfaces. Upon injection of the POPC vesicle solution into the flow chamber, the cantilevers underwent a surface stress change and deflection. Because the gold surface had previously been made inert by dithiolaromatic-PEG, the vesicles were only attracted to the negatively charged SiO2 surface of the microcantilevers and fused onto the SiO2 surface to form SLBs. This surface stress change is compressive, causing the cantilevers to bend away from the membrane-coated surface.30 The microcantilever bent with a deflection of 73 ± 6.0 nm. Thus, the measured surface stress, Δσads,SLB, for SLB formation on SiO2 was 17 ± 1.4 mN/m. Note that the vesicle solution flowed past the cantilever for 7 min (as indicated with the shaded area in Figure 2), after which, the valve was switched

gold surface of the cantilever. A blocking agent, a bovine casein (Alfa Aesar, Ward Hill, MA) solution, was used to prevent adsorption of the lipid bilayer and peptides on the silicon dioxide surface of the cantilever. An aqueous solution saturated with casein was filtered with a syringe filter with a 0.22-μm pore size to remove undissolved protein aggregates. A reference cantilever was functionalized with both the dithiolaromatic− PEG and casein. The microcantilever was used to probe the interactions between the peptide and the lipid membrane, as illustrated in Figure 1.

Figure 1. Schematic of the use of the lipid membrane-coated microcantilever to sense peptide adsorption and insertion. These interactions generate compressive stress on the microcantilevers and cause the microcantilever to bend toward the gold side.

Microcantilever Assay. The cantilever deflection results from the changes in the surface free energy associated with the physisorption or chemisorption of molecules to the cantilever surface. Preferential molecular adsorption on either of the cantilever surfaces induces a mismatch in the surface stress between the front and back surfaces of the cantilever, causing the cantilever to bend. The relationship between the cantilever deflection, Δz (m), and the change in surface stress, Δσ (N/ m), is described by Stoney’s equation:28 Δσ =

Et 2 ΔZ 3(1 − v)L2

Figure 2. PEP1 peptides at a concentration of 5 μM insert into POPC SLBs. Experimental cantilevers (solid lines) on the same chip are shown. The SiO2 surface of the reference cantilever (dashed line) is blocked by casein; thus, no lipid membrane adsorbs on it. The shaded areas indicate the time when the lipid vesicles (blue) or peptides (green) are introduced into the measurement chamber. Note that after the injection of the PEP1 solution, there is a slight surface stress change on the reference cantilever as a result of the minute amount of PEP1 adsorption on casein.

(1)

back to the PBS buffer port. Because the bilayer adsorption is irreversible over this time scale, the cantilevers remained deflected. The SLBs were stable for at least 2 h of observation time and did not desorb after the solution was switched to buffer. A control cantilever (dashed line in Figure 2), in which the gold and silicon surfaces were blocked with PEG and casein, respectively, showed the minimum surface stress change upon vesicle addition. PEP1 Binds to POPC SLBs at Low Concentrations. After the SLBs were prepared on the SiO2 surfaces of the microcantilevers, PEP1 was then introduced to the SLBs. Figure 2 compares the adsorption curve of PEP1 at a low concentration on the SLB-coated microcantilevers with that of a casein-coated cantilever. Upon PEP1 addition, there was an obvious increase in the surface stress, indicative of PEP1 insertion into the lipid bilayer. On the control microcantilever (dashed line), the surface stress remained small after the introduction of PEP1 to the surface, likely due to nonspecific binding of PEP1 peptides to the casein-coated surface through hydrophobic interactions. In addition, after the system was rinsed with buffer, no changes in surface stress were observed,

where ν is the Poisson’s ratio of the cantilever material, E the Young’s modulus, L the cantilever length, and t the cantilever thickness. The real-time deflection of the microcantilever was monitored using a Cantisens system from Concentris GmbH; this system utilizes a scanning laser diode aligned to the tip of the microcantilevers. The position of the reflected laser beam was captured using a position-sensitive detector (PSD) at a sampling frequency of 1 Hz. A solution of POPC vesicles was injected at 0.42 μL/s at 25 °C into the measurement chamber to form the SLB on the silicon dioxide surface of the microcantilever. A PBS solution containing PEP1 peptides at various concentrations was then injected to study the interaction between the peptide and the SLB. Small variations in the material properties of the cantilevers, such as the stiffness or thickness of the gold layer, resulted in different signals. Thus, the deflection of the microcantilevers was normalized by each cantilever’s thermomechanical sensitivity, using the change in deflection due to a 1 °C change in temperature.29 Each experiment was repeated at least three times on either the same or different chips, with a minimum of five cantilevers per chip. 10086

dx.doi.org/10.1021/ac501343b | Anal. Chem. 2014, 86, 10084−10090

Analytical Chemistry

Article

which indicates that the PEP1 adsorption is irreversible, unlike amphiphilic surfactants that have been studied previously.22 This is likely because the peptide system has a much stronger binding affinity for the membrane. High Concentrations of PEP1 Leads to POPC SLBs Lysis. The introduction of higher concentrations of PEP1 led to a precipitous drop in surface stress, indicative of the solubilization of the membrane from the sensor surface. We monitored the real-time surface process for the solubilization of an SLB when in contact with higher PEP1 concentrations. The surface stress dropped from 17 ± 1.4 mN/m to zero, suggesting that the SLB was perturbed and readily desorbed after contacting 26 μM PEP1, as shown in Figure 3. This disruption

Figure 4. (A) Surface stress generated on microcantilevers after the PEP1 solution is introduced to the SLBs. Three regions are separated by dashed lines. At concentrations 4 μM, the PEP1 peptides saturate the membrane surface, leading to a constant surface stress. At concentrations >26 μM, the membrane is solubilized, eventually leading to a bare SiO2 surface and zero surface stress. (B) Schematic representation of the PEP1 interaction with the SLB on a microcantilever surface. Four diagrams illustrate the possible conformations of the SLB before [0] and after [1−3] interacting with PEP1 at various concentrations.

Figure 3. POPC SLB is readily solubilized after contact with a high concentration of PEP1 peptides (26 μM). Both the experimental cantilevers (solid lines) on the same chip and the reference cantilever (dashed line, SiO2 surface is blocked by casein) are shown. The shaded areas indicate the time when lipid vesicles (blue) or peptides (green) are introduced to the measurement chamber. After the injection of the PEP1 solution, the adsorption of PEP1 on casein causes a surface stress change on the reference cantilever.

concentration in a linear fashion, indicating that the peptide adsorption was proportional to the concentration. At bulk concentrations higher than 4 μM, the surface stress did not measurably increase with the PEP1 concentration and remained at 25−28 mN/m, as shown in Figure 4A. The adsorbed PEP1 peptides saturated the lipid membrane at 4−5 μM bulk concentrations; thus, further increases in the peptide concentration did not lead to surface stress changes. At a sufficiently high concentration, the PEP1 peptides began to aggregate on the SLB surface, thus driving the interaction across the membrane, leading to pore formation.6 The lateral internal stress generated by these PEP1 aggregates was sufficient for membrane disruption13 and consequently pore formation in the SLB membrane, as shown in Figure 4B[2]. The stability and size of the pores are determined by the relative magnitudes of the membrane tension and the line tension at the edge of the pores, respectively.12 Other studies with lipid vesicles have shown the existence of a critical concentration when PEP1 interactions lead to vesicle swelling,5 which suggests pore formation.32,33 Although the inner leaflet in an SLB membrane was near a solid support, the PEP1 interaction with SLBs was similar to that which has been observed with free membranes. At PEP1 concentrations greater than 20 μM, the surface stress decreased. It is thought that after a sufficient number of peptides have aggregated and formed pores in the SLB, the internal stress of the membrane accumulates and reaches a value where it cannot be counterbalanced by the membrane

of the supported bilayer was also previously observed using AFM for PEP1 at concentrations larger than 26 μM.6 Note that for the control cantilever (dashed line) in Figure 3, the adsorption of PEP1 on casein is larger than that in Figure 2, because this adsorption is nonspecific and is mostly affected by the PEP1 concentration. Implication for the Mechanism of PEP1 Interaction with SLBs. PEP1 at various concentrations was introduced to POPC SLBs to study its interaction with lipid membranes. As shown in Figure 4A, this interaction was divided into three phases, based on the measured surface stress of the SLB. Figure 4B[0] illustrates the bilayer structure of the POPC lipids on the SiO2 surface before the addition of the PEP1 solution. At bulk PEP1 concentrations lower than 4 μM, the PEP1 peptides initially adsorbed to the membrane surfaces parallel to the membrane plane, because of their α-helical structure, which squeezed the membrane altitudinally,6 leading to additional lateral internal stress31 and incremental increases in the area of the outer leaflet of the membrane (Figure 4B[1]). The surfacebound PEP1 then continued to accumulate at higher concentrations in a Langmuir adsorption manner, further disrupting the lipid packing and thinning the membrane. At this concentration range, the surface stress quickly increased with 10087

dx.doi.org/10.1021/ac501343b | Anal. Chem. 2014, 86, 10084−10090

Analytical Chemistry

Article

tension; thus, the membrane solubilizes.13 The observed drop in surface stress indicates the removal of lipids from the cantilever surface, likely resulting in patchy SLBs (20 μM) and complete membrane solubilization at concentrations greater than 26 μM, as shown in Figure 4B[3]. It has been previously reported using QCM-D10 and SPR8 that lipid vesicles rupture upon PEP1 exposure at a concentration of ∼13 μM and that a supported membrane remains adsorbed on a gold or TiO2 surface. The PEP1-induced lysis of lipid vesicles is a vesicle sizedependent process in which a smaller vesicle can be more easily disrupted by the peptide.4 Compared with the PEP1-induced lysis of SLBs, the behavior of PEP1 when interacting with lipids is very likely related to the lipid curvature, which will be discussed in detail in subsequent sections. Surface Stress Analysis. The total surface area of the membrane−peptide complex expanded by the total area of the adsorbed peptides and resulted in an internal stress in the lipid membranes. SLBs can be elastically deformed; thus, an increase in the membrane surface area (or membrane strain in surface area) can be related to the internal stress by the membrane’s stretch modulus KA. The membrane internal stress, Δσmem, accounts for the lateral lipid−lipid and lipid-peptide interactions in the SLB.12,13 ⎛ A ⎞⎛ P ⎞ Δσmem = KA⎜ P ⎟⎜ ⎟ ⎝ AL ⎠⎝ L ⎠

There are interesting differences between free membranes and supported membranes. Mainly, the threshold concentration for causing vesicle rupture is lower than that for inducing SLB solubilization. In addition, for the same peptide concentration, smaller vesicles are easier to disrupt.4 This result indicates that the PEP1−lipid interaction is likely dependent on membrane curvature. It has also been reported that amphipathic α-helical peptides can be used as a sensing motif for membrane curvature.36,37 A membrane with higher curvature possesses more binding sites38 and thus facilitates pore formation and lipid lysis.39,40 Supported lipid membranes, with their very small curvatures, need a much higher critical concentration of PEP1 peptides to induce membrane solubilization. In addition, the underlying solid support hinders lipid mobility, making solubilization more difficult. Therefore, we hypothesize that the PEP1 interactions with different membrane models, either lipid vesicles or SLBs, are governed by the same process, in which the peptide concentration and membrane curvature synergistically determine the adsorption, destabilization, or solubilization of the lipid membrane. We previously studied the interaction of amphipathic surfactants, lysolipids, with SLBs; this interaction also involves lysolipid adsorption and eventual membrane solubilization.22 Although lysolipids behave similarly to the PEP1 peptide when interacting with SLB, the PEP1−SLB interaction is stronger, because the adsorbed PEP1 does not desorb from the SLB after a buffer rinse while the lysolipid desorbs easily. In addition, PEP1 disrupts the SLBs at a much lower concentration and is thus more efficient at solubilizing the lipid bilayers and removing them from the solid surface, compared with the lysolipids. The lipid interaction with PEP1 is also comparable to that with antimicrobial peptides, because they both have amphipathic α-helical and membrane-active characteristics. The membrane interaction with antimicrobial peptides is determined by the molar peptide-to-lipid ratio (P/L) in a two-state model.41 The peptides bind to the membrane at a ratio lower than the critical P/L but cause pore formation when the ratio is higher than the critical P/L. With the presence of pores, additional peptides do not give rise to additional membrane thinning effects, and thus, no further internal stress is generated.12 This observation also explains the plateau in Figure 4A, where the surface stress does not increase with PEP1 concentration. The amphipathic α-helical property is essential for the interaction of the peptides with the lipid membrane; thus, the model developed for antimicrobial peptides can be generally applied to the PEP1−SLB interaction.6,32 Because the peptide concentration is proportionally related to P/L, the mechanism of PEP1−SLB interaction can be presented by a three-state model: at a P/L value below the first critical ratio (P/L < P/L*), PEP1 peptides adsorb onto the SLB in parallel; at a P/L value between the two critical ratios (P/L* < P/L < P/ L**), PEP1 peptides insert into the SLB and form pores across the membrane; and at P/L values above the second critical ratio (P/L > P/L**), PEP1 peptides solubilize the SLB, and the lipids are removed from the surface. The main difference between the two systems is that the critical P/L for PEP1 is lower than that for most antimicrobial peptides, indicating the efficiency with which the PEP1 peptide can disrupt lipid membranes.39 In addition, comparing the response of the PEP1 adsorption onto POPC SLBs using the microcantilever sensors with that using other sensors, such as QCM-D, indicates that the partitioning of the peptide into an SLB is more responsive to

(2)

where AP and AL represent the molecular area of the peptide and the lipid, respectively. P/L is the number ratio of membrane-bound peptide to lipid. Another source of surface stress, Δσads, results from lipid-surface interactions, which include the chemical potential gained from adsorption. Thus, the total surface stress, Δσtotal, measured on the microcantilevers is composed of the contributions from lipid−lipid and lipid−peptide interactions (Δσmem) and lipid−surface interactions (Δσads): ⎛A ⎞ Δσtotal = Δσads + Δσmem = Δσads + KA ⎜ P ⎟ ⎝ AL ⎠

(3)

The internal stress of the membrane is estimated from eq 2 by taking KA to be 240 mN/m for a POPC membrane,34 AP be 360 Å2 and AL be 65 Å2 for lipid membranes in a fluid state.6 Huang et al. predicted that, at the concentrations when pore formation occurs, the typical value for Δσmem is 5−15 mN/m,32 which is close to the surface stress measurement of ∼13 mN/m using the microcantilevers. Exclusion of the PEP1 adsorption term (Δσads) and only accounting for the membrane internal stress term (Δσmem) for the calculation of the total measured surface stress allows for an estimation of the P/L value of 1/ 102, which indicates that one PEP1 is surrounded by 102 lipid molecules. The actual P/L ratio may be smaller because the adsorption term (Δσads) is not taken into consideration.

■

DISCUSSION The interaction of PEP1 peptides with lipid vesicles has been investigated and is currently used to rupture lipid vesicles and form a supported membrane under unfavorable conditions.8,10,35 The mechanism is as follows: PEP1 peptides first bind to the lipid membrane, induce pore formation, which leads to vesicle swelling, and eventually rupture the vesicles.5 Our results show that this mechanism also occurs on SLBs: the peptides adsorb, destabilize, and solubilize the lipid bilayer. 10088

dx.doi.org/10.1021/ac501343b | Anal. Chem. 2014, 86, 10084−10090

Analytical Chemistry

Article

(4) Cho, N.-J.; Dvory-Sobol, H.; Xiong, A.; Cho, S.-J.; Frank, C. W.; Glenn, J. S. ACS Chem. Biol. 2009, 4, 1061−1067. (5) Cho, N.-J.; Wang, G.; Edvardsson, M.; Glenn, J. S.; Hook, F.; Frank, C. W. Anal. Chem. 2009, 81, 4752−4761. (6) Cho, N.-J.; Cho, S.-J.; Hardesty, J. O.; Glenn, J. S.; Frank, C. W. Langmuir 2007, 23, 10855−10863. (7) Cho, N.-J.; Cheong, K. H.; Lee, C.; Frank, C. W.; Glenn, J. S. J. Virol. 2007, 81, 6682−6689. (8) Chah, S.; Zare, R. N. Phys. Chem. Chem. Phys. 2008, 10, 3203− 3208. (9) Ludtke, S. J.; He, K.; Heller, W. T.; Harroun, T. A.; Yang, L.; Huang, H. W. Biochemistry 1996, 35, 13723−13728. (10) Cho, N.-J.; Cho, S.-J.; Cheong, K. H.; Glenn, J. S.; Frank, C. W. J. Am. Chem. Soc. 2007, 129, 10050−10051. (11) Bobardt, M. D.; Cheng, G.; de Witte, L.; Selvarajah, S.; Chatterji, U.; Sanders-Beer, B. E.; Geijtenbeek, T. B. H.; Chisari, F. V.; Gallay, P. A. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 5525−5530. (12) Huang, H. W.; Chen, F. Y.; Lee, M. T. Phys. Rev. Lett. 2004, 92, 198304. (13) Lee, M. T.; Chen, F. Y.; Huang, H. W. Biochemistry 2004, 43, 3590−3599. (14) Lee, M.-T.; Hung, W.-C.; Chen, F.-Y.; Huang, H. W. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 5087−5092. (15) van den Bogaart, G.; Guzman, J. V.; Mika, J. T.; Poolman, B. J. Biol. Chem. 2008, 283, 33854−33857. (16) Raiteri, R.; Grattarola, M.; Butt, H.; Skládal, P. Sens. Actuators B 2001, 79, 115−26. (17) Braun, T.; Ghatkesar, M. K.; Backmann, N.; Grange, W.; Boulanger, P.; Letellier, L.; Lang, H.; Bietsch, A.; Gerber, C.; Hegner, M. Nat. Nanotechnol. 2009, 3, 179−85. (18) Knowles, T. P. J.; Shu, W.; Huber, F.; Lang, H. P.; Gerber, C.; Dobson, C. M.; Welland, M. E. Nanotechnology 2008, 19, 384007. (19) White, D. A.; Buell, A. K.; Dobson, C. M.; Welland, M. E.; Knowles, T. P. J. FEBS Lett. 2009, 583, 2587−2592. (20) Ghatkesar, M. K.; Lang, H.-P.; Gerber, C.; Hegner, M.; Braun, T. PLoS One 2008, 3, e3610. (21) Liu, K.-W.; Biswal, S. L. Anal. Chem. 2010, 82, 7527−7532. (22) Liu, K.-W.; Biswal, S. L. Anal. Chem. 2011, 83, 4794−4801. (23) Wang, J.; Segatori, L.; Biswal, S. L. Soft Matter 2014, 10, 6417− 6424. (24) Hardy, G. J.; Nayak, R.; Zauscher, S. Curr. Opin. Colloid Interface Sci. 2013, 18, 448−58. (25) Hope, M. J.; Bally, M. B.; Webb, G.; Cullis, P. R. Biochim. Biophys. Acta 1985, 812, 55−65. (26) Mecke, A.; Lee, D. K.; Ramamoorthy, A.; Orr, B. G.; Holl, M. M. B. Biophys. J. 2005, 89, 4043−4050. (27) Sato, H.; Felix, J. B. Biochim. Biophys. Acta, Biomembr. 2006, 1758, 1245−1256. (28) Stoney, G. G. Proc. R. Soc. London, Ser. A 1909, 82, 172−175. (29) Wang, J.; Morton, M. J.; Elliott, C. T.; Karoonuthaisiri, N.; Segatori, L.; Biswal, S. L. Anal. Chem. 2014, 86, 1671−1678. (30) Wang, J.; Liu, K.-W.; Segatori, L.; Biswal, S. L. J. Phys. Chem. B 2014, 118, 171−178. (31) Reddy, A. S.; Warshaviak, D. T.; Chachisvilis, M. Biochim. Biophys. Acta, Biomembr. 2012, 1818, 2271−2281. (32) Huang, H. W. Biochim. Biophys. Acta, Biomembr. 2006, 1758, 1292−1302. (33) Longo, M. L.; Waring, A. J.; Gordon, L. M.; Hammer, D. A. Langmuir 1998, 14, 2385−2395. (34) Rawicz, W.; Olbrich, K. C.; McIntosh, T.; Needham, D.; Evans, E. Biophys. J. 2000, 79, 328−339. (35) Hardy, G. J.; Nayak, R.; Munir Alam, S.; Shapter, J. G.; Heinrich, F.; Zauscher, S. J. Mater. Chem. 2012, 22, 19506−19513. (36) Antonny, B. In Annual Review of Biochemistry, Vol. 80; Kornberg, R. D., Raetz, C. R. H., Rothman, J. E., Thorner, J. W., Eds.; Annual Reviews: Palo Alto, CA, 2011; pp 101−123. (37) Chen, R.; Mark, A. E. Eur. Biophys. J. Biophys. Lett. 2011, 40, 545−553.

surface stress changes than to mass changes. For the adsorption of PEP1 to lipid membranes at PEP1 concentrations higher than the critical concentration, the frequency change measured on QCM-D will be estimated to be ∼3.5 Hz, while the frequency change for the SLB formation is ∼25 Hz.7 On microcantilevers, the maximum surface stress change induced by PEP1 is 13 mN/m, while that induced by the SLB formation is 17 mN/m. Thus, the microcantilever sensor is more sensitive to the adsorption of small molecules to membranes and to interactions that are involved with a change in the internal lateral stress of the membrane.

■

CONCLUSION The bending motion of lipid membrane-coated microcantilevers is a direct surface measurement of the surface stress change as PEP1 peptides bind to the membrane. Interactions between PEP1 and the lipid membranes are measured at different peptide concentrations. From the plot of surface stress versus PEP1 concentration, three phases of interactions are observed. At concentrations P/L**), the surface stress decreases, indicating both damage to the membrane integrity and membrane solubilization. Compared with other surface-sensitive tools, such as QCM-D7 and SPR,8 the microcantilever sensor is much more sensitive at probing the adsorption and interactions of small molecules, because the change in the internal stress of the membrane can also be detected as a surface stress change. Thus, the lipid membranecoated microcantilever sensor is capable of characterizing the kinetics and dynamics of membrane-peptide interactions with high sensitivity.

■

ASSOCIATED CONTENT

* Supporting Information S

Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Tel.: 713-348-6055. E-mail: [email protected]. Author Contributions †

These authors contributed equally to this work.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This research was funded in part by the Welch Foundation (Grant No. C-1755) to S.L.B. and the Riki Kobayashi Fellowship to J.W.

■

REFERENCES

(1) Di Bisceglie, A. M. Hepatology 2000, 31, 1014−1018. (2) Cheng, G.; Montero, A.; Gastaminza, P.; Whitten-Bauer, C.; Wieland, S. F.; Isogawa, M.; Fredericksen, B.; Selvarajah, S.; Gallay, P. A.; Ghadiri, M. R.; Chisari, F. V. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 3088−3093. (3) Elazar, M.; Cheong, K. H.; Liu, P.; Greenberg, H. B.; Rice, C. M.; Glenn, J. S. J. Virol. 2003, 77, 6055−6061. 10089

dx.doi.org/10.1021/ac501343b | Anal. Chem. 2014, 86, 10084−10090

Analytical Chemistry

Article

(38) Hatzakis, N. S.; Bhatia, V. K.; Larsen, J.; Madsen, K. L.; Bolinger, P.-Y.; Kunding, A. H.; Castillo, J.; Gether, U.; Hedegard, P.; Stamou, D. Nat. Chem. Biol. 2009, 5, 835−841. (39) Tabaei, S. R.; Rabe, M.; Zhdanov, V. P.; Cho, N.-J.; Höök, F. Nano Lett. 2012, 12, 5719−5725. (40) Jackman, J. A.; Zan, G. H.; Zhdanov, V. P.; Cho, N.-J. J. Phys. Chem. B 2013, 117, 16117−16128. (41) Huang, H. W. Biochemistry 2000, 39, 8347−8352.

10090

dx.doi.org/10.1021/ac501343b | Anal. Chem. 2014, 86, 10084−10090