5022
Langmuir 2007, 23, 5022-5028
Adsorption of the Fusogenic Peptide B18 onto Solid Surfaces: Insights into the Mechanism of Peptide Assembly Sandra Rocha,*,†,‡ M. Carmo Pereira,‡ Manuel A. N. Coelho,‡ Helmuth Mo¨hwald,† and Gerald Brezesinski† Max Planck Institute of Colloids and Interfaces, Research Campus Golm, Am Mu¨hlenberg 1, D-14476 Potsdam, Germany, and LEPAE, Department of Chemical Engineering, Faculty of Engineering, UniVersity of Porto, Rua Dr. Roberto Frias, 4200-465 Porto, Portugal ReceiVed September 25, 2006. In Final Form: January 16, 2007 The adsorption and assembly of B18 peptide on various solid surfaces were studied by reflectometry techniques and atomic force microscopy. B18 is the minimal membrane binding and fusogenic motif of the sea urchin protein bindin, which mediates the fertilization process. Silicon substrates were modified to obtain hydrophilic charged surfaces (oxide layer and polyelectrolyte multilayers) and hydrophobic surfaces (octadecyltrichlorosilane). B18 does not adsorb on hydrophilic positively charged surfaces, which was attributed to electrostatic repulsion since the peptide is positively charged. In contrast, the peptide irreversibly adsorbs on negatively charged hydrophilic as well as on hydrophobic surfaces. B18 showed higher affinity for hydrophobic surfaces than for hydrophilic negatively charged surfaces, which must be due to the presence of hydrophobic side chains at both ends of the molecule. Atomic force microscopy provided the indication that lateral diffusion on the surface affects the adsorption process of B18 on hydrophobic surfaces. The adsorption of the peptide on negatively charged surfaces was characterized by the formation of globular clusters.
Introduction Many biological processes such as membrane fusion are dictated by interfacial phenomena. Fusion is essential in numerous intra and intercellular processes such as fertilization, vesicular trafficking, muscle development, and enveloped virus infections. In biological cells fusion of membranes is induced by proteins.1 During fusion, proteins are required to overcome the repulsive surface forces such as hydration and electrostatic interactions as well as steric barriers that act against the close approach of the opposing membranes.2 Fusion proteins contain a short hydrophobic sequence of 20-25 amino acid residues, known as fusion peptide, which interacts directly with the membrane,3 and are therefore functional at interfaces. Surface interactions may be responsible for protein structure changes, which will determine its function. These interactions may as well induce aggregation into for instance amyloid β-structures, as it was reported for B18 peptide and several virus fusion peptides.4,5 B18 peptide (LGLLLRHLRHHSNLLANI) is the amino acid sequence 103 to 120 of the membrane-associated acrosomal protein bindin, which plays a key role in the fertilization process of Strongylocentrotus purpuratus (purple sea urchin).6 The sequence B18 is recognized as the minimal membrane binding and fusogenic motif.6,7 These 18 amino acids are perfectly conserved among all known sea urchin species. * Corresponding author. † Max Planck Institute of Colloids and Interfaces. ‡ University of Porto. (1) Sollner, T.; Whiteheart, S. W.; Brunner, M.; Erdjument-Bromage, H.; Geromanos, S.; Tempst, P.; Rothman, J. E. Nature 1993, 362, 318-324. (2) Israelachvili, J. N. Intermolecular and Surface Forces, 2nd ed.; Academic Press: London, 1992; Chapter 18. (3) Pe´cheur, E. I.; Sainte-Marie, J.; Bienvenu¨e, A.; Hoekstra, D. J. Membr. Biol. 1999, 167, 1-17. (4) Ulrich, A. S.; Tichelaar, W.; Fo¨rster, G.; Zscho¨rnig, O.; Weinkauf, S.; Meyer, H. W. Biophys. J. 1999, 77, 829-841. (5) Tamm, K. L.; Han, X. Biosci. Rep. 2000, 20, 501-518. (6) Miraglia, S. J.; Glabe, C. G. Biochem. Biophys. Acta 1993, 1145, 191198. (7) Ulrich, A. S.; Otter, M.; Glabe, C. G.; Hoekstra, D. J. Biol. Chem. 1998, 273, 16748-16755.
The B18 conformation is depending on solution conditions. At slightly acidic pH, the peptide shows a random coil conformation, whereas at neutral pH, the sequence has a strong tendency to self-assemble.4,8 At high concentration and pH above 7, the peptide forms fibrils, consisting of twisted ribbons that are assembled from three to five protofilaments with widths of about 5 nm each.4 The fibrils have a cross β-sheet conformation and stain positively to Congo red dye, similar to amyloid fibrils. There is a homology in the amino acid composition of fusion peptides and the amyloidogenic peptides prion fragments and amyloid β-peptide.9 Such as the fusion peptides, prion fragments and amyloid β-peptide have a first part rich in hydrophobic amino acids whereas the second one contains more small amino acid residues (alanine, glycine).9 Fusogenic properties were also described for these peptides and are thought to contribute to their neurotoxicity by destabilizing membranes.10,11 B18 undergoes a transition from coil to helix structure in trifluoroethanoll8 and in the presence of fluorinated particles.12 However, in trifluoroethanol at high peptide concentrations and pH 7.5, an intermediate state constituted by oligomeric species appears to exist. Transition metal ions bind to B18, leading as well to the formation of aggregates. Electron microscopy assay showed that the complexes of B18 with zinc and cupper consist of smooth globular clusters.4 B18 peptide is able to induce aggregation and fusion of lipid vesicles, for what it requires zinc ions.4,7,13 At low peptide to (8) Glaser, R. W.; Gru¨ne, M.; Wandelt, C.; Ulrich, A. S. Biochemistry 1999, 38, 2560-2569. (9) Angel, V. D.; Dupuis, F.; Mornon, J. P.; Callebaut, I. Biochem. Biophys. Res. Commun. 2002, 293, 1153-1160. (10) Pillot, T.; Goethals, M.; Vanloo, B.; Talussot, C.; Brasseur, R.; Vandekerckhove, J.; Rosseneu, M.; Lins, L. J. Biol. Chem. 1996, 271, 2875728765. (11) Pillot, T.; Lins, L.; Goethals, M.; Vanloo, B.; Baert, J.; Vandekerckhove, J.; Rosseneu, M.; Brasseur, R. J. Mol. Biol. 1997, 274, 381-393. (12) Rocha, S.; Thu¨nemann, A. F.; Pereira, M. C.; Coelho, M. A. N.; Mo¨hwald, H.; Brezesinski, G. ChemBioChem 2005, 6, 1-4. (13) Binder, H.; Arnold, K.; Ulrich, A. S.; Zscho¨rnig, O. Biochim. Biophys. Acta 2000, 1468, 345-358.
10.1021/la0628120 CCC: $37.00 © 2007 American Chemical Society Published on Web 03/29/2007
Fusogenic Peptide B18
lipid ratio, B18 adopts an R-helix structure at zwitterionic and negatively charged membranes. At high ratio, however, solidstate NMR spectroscopy and electron microscopy studies revealed an oligomeric β-sheet structure14 and fibril formation.4 Despite the structural studies, the molecular mechanism of B18 binding to membranes and of the aggregate formation has not been completely elucidated. Because B18 induces the fusion of membranes, interfacial phenomena and interaction between peptide and surfaces may be involved in the formation of peptide aggregates. The adsorption of the peptide on hydrophilic (oxide layer and polyelectrolyte multilayers) and hydrophobic (octadecyltrichlorosilane) solid surfaces without specific interactions, which have to be considered on membrane surfaces, was studied by neutron and X-ray reflectivity experiments and correlated with aggregate formation. Atomic force microscopy experiments focus on the morphology of B18 aggregates deposited on such surfaces as a function of time. Experimental Section Materials. B18 peptide (Mw ) 2090 g/mol) was kindly provided by Olaf Zscho¨rnig (Institute for Medical Physics and Biophysics, University of Leipzig, Germany). The peptide was synthesized by standard solid-phase Fmoc protocols and purified by reverse phase high-pressure liquid chromatography, as described elsewhere.4,7 As a result of the synthesis, B18 peptide is blocked at the C-terminus with an amide group. Peptide stock solutions were prepared by dissolving B18 in double distilled water, where the peptide is fully soluble and does not self-aggregate. The solution was then diluted to a concentration of 4.8 µM with 10 mM hepes buffer (99.5%, Sigma-Aldrich) in order to obtain a pH of 7.5. Sodium azide (0.05 mM, Fluka) was added to prevent microbial growth. The peptide was initially in random coil conformation as determined by circular dichroism spectroscopy (Jasco spectropolarimeter J-715). Sodium poly(styrene sulfonate) (Mw ) 70 000 g/mol), poly(allylamine hydrochloride) (Mw ) 50000-65000 g/mol), poly(ethyleneimine) (Mw ) 750 000 g/ mol, and deuterium oxide (D2O) (99.0% isotope enrichment) were supplied by Sigma-Aldrich Co (Germany). Toluene (99.7%), isopropanol (99.7%) and n-octadecyltrichlorosilane (98%) were from Merck (Darmstadt, Germany). All commercial chemicals were used without further purification. The water was deionized using a Purelab Plus UV/UF (Elga LabWater VivendiWater Systems Deutschland Holding GmbH) purification system and had a conductivity of 0.055 µS cm-1. Surface Modification. The silicon wafers (Silchem Handelsgesellschaft GmbH Freiberg, Germany) and the silicon blocks (prepared by Andrea Holm, Tann/Ndb., Germany) were first cleaned by the RCA method.15 The substrates were immersed in a mixture of H2O, H2O2 (pa, 30% aqueous solution, Fluka), and NH4OH (pa, 28% aqueous solution, Roth) with a volume ratio of 5:1:1, which was heated to 80 °C for 10 min. Afterward the wafers were rinsed extensively and stored in water for not longer than 1 day before coating. The polished surfaces of silicon substrates bear a native oxide layer. The oxide layer surface consists of siloxane bonds (Si-OSi), which rapidly acquire silanol groups (Si-OH) at the surface from contact with water or atmospheric moisture. After the RCA procedure, the silicon oxide surface is well saturated with silanol groups, conferring very high hydrophilicity. Above pH 2, the oxide surface is negatively charged and the charge density is almost constant over pH 3-8.16 The charged multilayers were prepared by deposition of oppositely charged polyelectrolytes (10-2 monomer mol/L) using the layerby-layer self-assembly technique.17 All layers were assembled in the presence of 1 M NaCl. The silicon block was initially immersed (14) Barre´, P.; Zscho¨rnig, O.; Arnold, K.; Huster, D. Biochemistry 2003, 42, 8377-8386. (15) Kern, W.; Puotinen, D. A. RCA Ver. 1970, 31, 187-206. (16) Iler, R. K. The Chemistry of Silica, Wiley: New York, 1979. (17) Decher, G. Science 1997, 277, 1232-1237.
Langmuir, Vol. 23, No. 9, 2007 5023 in a solution of poly(ethyleneimine) (PEI) for 20 min, and then washed with water to remove the excess of polymer. Then the negatively charged polyelectrolyte sodium poly(styrene sulfonate) (PSS) and the positively charged polyelectrolyte poly(allylamine hydrochloride) (PAH) were alternately deposited by immersing the substrate for 20 min in the corresponding polyelectrolyte solution and rinsing with water between the adsorption steps. After deposition of the last layer the sample was dried with a gentle nitrogen stream. The films PEI(PSS/PAH)6 and PEI(PSS/PAH)6PSS were obtained. The contact angles amount to 45° for multilayers with PSS as the outermost layer, and 65° for PAH. The contact angles were determined as described by Wong et al.18 Silanization of silicon substrates was attained with n-octadecyltrichlorosilane (OTS) dissolved in toluene at a concentration of 2.5 × 10-3 mol/L. The substrates were first dried with nitrogen gas, as well as the surrounding atmosphere to eliminate the water. The substrates were then immersed in the solution overnight at a temperature of 4 °C. The excess of OTS was removed by washing the substrates with toluene and isopropanol alternately several times. The contact angle of the OTS layer is 105°. Neutron and X-Ray Reflectivity Experiments. Neutron reflectometry measurements were performed at the Hahn-Meitner Institute, Berlin, on the V6 reflectometer with a θ/2θ geometry. The experiments were performed in a solid/liquid experimental cell, which consisted of a single silicon block with size of 15 × 80 × 50 mm3 and a Teflon cell that was fixed to the bottom of the block. A detailed description of the instrument is given elsewhere.19 A neutron wavelength of 4.66 Å was selected by a graphite monochromator in the incident white beam. The resolution was set by a slit system on the incident side to ∆q ) 0.001 Å-1 for q e 0.08 Å-1, and ∆q ) 0.002 Å-1 for larger q values. A beam of rectangular cross section 0.5 × 40 mm2 up to q ) 0.08 Å-1 and 1 × 40 mm2 for higher q values impinged on the samples through the silicon block. The scattered neutrons were recorded with a 3He-detector in single θ/2θ steps with a complete run from 0.0047 to 0.12 Å-1 consuming typically 5-7 h of beamtime. The off-specular signal was collected simultaneously in a 3He counter shifted from the specular position by 0.44° toward larger 2θ value. The reflectometer has a vertical scattering geometry. The measured intensity in the off-specular channel was directly subtracted from the intensity in the specular channel to obtain the background corrected intensity. If the adsorbed material (p) forms a uniform layer containing water, the scattering length density (SLD) Fexp, obtained from the experiment, is a weighted sum of the SLD of the material and the SLD of water (w) present in the layer: Fexp ) Fpφp + (1 - φp) Fw
(1)
where φp is the volume fraction of material p in the layer. Because of the large difference between the scattering length density of D2O and that of the organic layer the determination of the water content in the adsorption layer is possible. Specular X-ray reflectometry was measured using a θ (reflection angle)/2θ (detector angle) configuration of a θ/θ STOE reflectometer and using a conventional X-ray tube source. X-rays of 1.54 Å wavelength (Cu KR radiation) were selected by a graphite secondary monochromator complemented with electronic discrimination (scintillation counter). The X-ray beam was collimated by a set of adjustable slits with micrometer precision. The divergence of the incoming beam was 0.1° and the 2θ resolution was 0.05°. All measurements were carried out with a voltage of 30 kV and a current of 15 mA. The samples were scanned over the 2θ region of 0.2° to 5°. The reflected intensity in X-ray and neutron experiments was measured as a function of detector angle, θ, where θ describes the incident (or scattering) angle in θ/2θ geometry, and subsequently plotted versus q, the momentum transferred perpendicularly to the sample, which is defined by (4π/λ) sin θ. The measured intensity (18) Wong, J. E.; Florian Rehfeldt, F.; Ha¨nni, P.; Tanaka, M.; Klitzing, R. Macromolecules 2004, 37, 7285-7289. (19) Mezei, F.; Goloub, R.; Klose, F.; Toews, H. Phys. B 1995, 213&214, 898-900.
5024 Langmuir, Vol. 23, No. 9, 2007
Rocha et al.
was footprint corrected and normalized to the measured incident intensity I0 to obtain the reflectivity, R(q), of the interface. The spectra were analyzed by applying the standard fitting routine Parratt 32.20 The calculation is based on Parratt’s recursion scheme for stratified media. The film was modeled as consisting of layers of specific thickness and scattering length density or electron density. The model reflectivity profile calculated using the dynamic iterative model was compared to the measured one. Then the model was adjusted to the best chi-square fit to the data. The thickness, scattering length density (neutron) or electron density (X-ray) and roughness of each layer were extracted from box model fits. Atomic Force Microscopy. The wafers were removed at each time point from one of the incubation vials for atomic force microscopy (AFM) measurements. The samples were washed with water to remove unbound peptide and dried with nitrogen gas. AFM imaging was performed in air in tapping mode with a Nanoscope III Multi-mode instrument. Measurements were carried out using silicon cantilevers (Nanoworld, Switzerland) with a resonance frequency of 290-350 kHz. Digital resolution was 512 × 512 pixels. At least three regions of the surface were examined to verify that similar morphologies existed throughout the sample. The images were analyzed using the Nanoscope III software. Height sizes were estimated by section analysis and the root-mean-square roughness was calculated as
σ)
∑ (Z - Zh )
x
Figure 1. Neutron reflectivity profile of positively charged PEI(PSS/PAH)6 multilayers in D2O: (1) bare film; (2) film exposed to B18 solution for 12 h. The curves are offset for clarity.
2
i
N
where Zi is the height value for the ith pixel on the AFM image, Z h is the average on Zi, and N is the number of points in the selected area.
Results Adsorption on Hydrophilic Surfaces.The adsorption of B18 peptide on charged PEI(PSS/PAH)6 and PEI(PSS/PAH)6PSS films was characterized by neutron reflectometry. The technique allows in situ solid/liquid measurements which are suitable for characterizing films in water. The polyelectrolyte multilayers have affinity toward water and are sensitive to humidity,18 and consequently their thickness will vary according to the water content and to the atmospheric humidity. In this case it is necessary to precisely control the humidity or perform the measurements in water so the changes detected in the film after contact with B18 will be due to peptide adsorption and not due to water content variation. The reflectivity profile of the pure films in D2O was first measured. The subphase was then exchanged by the peptide solution. After 12 h adsorption time, the peptide solution was exchanged by pure D2O, and neutron reflectivity measurements were repeated. Figure 1 shows the neutron reflectivity profile of the positively charged PEI(PSS/PAH)6 film before (curve 1) and after exposure to B18 peptide solution (curve 2). No differences in the position and amplitude of the Kiessig fringes were observed in the reflectivity curves, indicating that the thickness and composition of the film were unchanged. Hence, adsorption and penetration of B18 can be excluded for the positively charged film. The effect of B18 peptide on the negatively charged PEI(PSS/PAH)6PSS film can be seen in Figure 2. B18 exposure resulted in a clear shift of the position of the minima to smaller q values, indicating an increase in the film thickness. The structural parameters of the pure PEI(PSS/PAH)6PSS film were obtained by fitting a two-layer model to the reflectivity profile, the first (20) Braun, C. Parratt32 Fitting Routine for ReflectiVity Data; HMI: Berlin, 1997.
Figure 2. Neutron reflectivity profile of negatively charged PEI(PSS/PAH)6PSS multilayers in D2O: (1) bare film; (2) film exposed to B18 solution for 12 h. The continuous lines are profiles calculated using parameters of Table 1. The curves are offset for clarity.
layer representing the silicon oxide and the second layer the polymer film. The thickness and the scattering length density of the polymeric film were estimated to be 358 Å and 3.79 × 10-6 Å- 2, respectively. The model used for fitting the reflectivity data after adsorption of B18 to PSS-terminated multilayers was the same two layer model as used beforehand where the second layer is now describing the polymer film plus adsorbed B18. A thickness of 369 Å was obtained. Therefore B18 adsorption resulted in a thickness increase of 11 Å. The fitting parameters are resumed in Table 1. Atomic force microscopy (AFM) imaging of PEI(PSS/ PAH)6PSS after B18 peptide adsorption revealed small differences when compared to the bare surface (data not shown). Nevertheless, by the neutron reflectivity data it is assumed that the peptide adsorbs in the form of patches, which are placed between the polymer aggregates since adsorbed B18 could not be fitted as an additional single uniform layer and the roughness was large. The adsorption of B18 on silicon oxide layer was characterized by AFM. Reflectivity measurements of these samples were not performed since both the silicon oxide layer (in the order of 10 Å) and the peptide adsorption layer are very thin. For thin layers, structural information can be extracted only at high q values. However, at high q values (>0.2 Å-1), the reflectivity will rapidly
Fusogenic Peptide B18
Langmuir, Vol. 23, No. 9, 2007 5025
Table 1: Structural Parameters Used to Fit the Reflectivity Profiles for the System PEI(PSS/PAH)6PSS and B18 Peptidea layer silicon oxide before adsorption film D2O after adsorption film and B18 D2O
d (Å) 8(2
F (10-6 Å-2) 3.41 ( 0.10
σ (Å) 5(1
358 ( 1 -
3.79 ( 0.10 6.11 ( 0.20(a)
9(3 23 ( 4
369 ( 1 -
3.79 ( 0.10 6.11 ( 0.20
2(1 25 ( 1
a d, F and σ are the thickness, scattering length density and roughness, respectively. The deviation from the expected value (6.37 × 10-6 Å-2) is due to traces of H2O.
decrease to the background level (due to incoherent scattering from the solvent). For q < 0.2 Å- 1 the reflectivity profile of a thin silicon oxide layer decays monotonically without the appearance of Kiessig fringes. As the peptide layer is expected to be in the range of 10-20 Å, changes in the overall thickness cannot be easily detected with the instruments used in this work. The AFM images of the initial bare silicon surface after incubation in buffer solution are shown in Figure 3. As expected, the substrate surface is flat and featureless (Figure 3a). Thus, small changes in the morphology of the surface upon peptide adsorption are presumably easy to detect. At early times after exposing the silicon substrate to the peptide solution, no significant differences were found on the surface when compared to the bare substrate (Figure 3b). After increasing the incubation time to 12 h, spherical structures closely packed are observed (Figure 3c). However, areas without adsorbed molecules are also seen. After 24 h incubation time, the color of the structures became brighter, indicating that their heights increased (Figure 3d). Nonetheless areas without adsorbed molecules are still observed. The heights of the spherical structures were quantified as 13 ( 2 Å after 12 h adsorption time, which matches the thickness of B18 adsorption layer on PEI(PSS/ PAH)6PSS film as determined by neutron reflectivity. This implies that there was no water inside the adsorbed layer, except between the individual B18 patches. From this finding it is possible to determine the volume fraction of D2O in the peptide layer adsorbed on PSS terminated film. Since the peptide contains hydrogens on the backbone and amino acid side groups, which readily exchange with D2O, the effect of H/D exchange on the total scattering length of the peptide has to be considered. The exchangeable protons predicted for B18 in its nonionized state include 17 backbone amides, 2 N-terminal protons (C-terminal is blocked) and 15 labile side chain protons. The estimation of the B18 molecular volume was based in partial specific volumes given by Harpaz et al.,21 leading to a value of 2529 Å3. The F of the peptide with 34 exchanged hydrogens was subsequently calculated to be 3.15 × 10-6 Å-2 using tabulated data for the scattering length of individual atoms.22 This leads to a volume fraction of peptide of 0.78 and 0.22 of D2O. The accuracy of this value is affected by the estimation of the peptide molecular volume, because the molecular volume might be different when the peptide is adsorbed at an interface. After 24 h adsorption time, two height distributions were found, 16 ( 1 Å and 23 ( 2 Å on the silicon surface, indicating that molecules are adsorbing on already adsorbed peptide. The roughness values were low ranging from 4.0 ( 0.2 Å after 12 h to 5.7 ( 0.4 Å after 24 h, which are close to the values found for the bare silicon substrate (3.2 ( 0.1) Å. Adsorption on Hydrophobic Surface. X-ray measurements were carried out to characterize the adsorption behavior of B18 (21) Harpaz, Y.; Gerstein, M.; Chothia C. Structure 1994 2, 641-649.
on hydrophobic surfaces. Octadecyltrichlorosilane (OTS) forms thin layers in the order of 25 Å.23 After a defined adsorption period, the samples were rinsed with ultrapure water and dried in a nitrogen flow. Figure 4 shows the reflectivity curve of the pure OTS layer (curve 1), which was fitted with a two-layer model, the first layer corresponding to the silicon oxide and the second layer to OTS. The electron density and the thickness of the bare OTS layer were 0.24 e-/Å3 and 24 Å, respectively. The reflectivity profile of the hydrophobic surface changed after incubation in B18 peptide solution. After 1 h adsorption time, the reflectivity curve showed a pronounced shift of the position of the first minimum to smaller q values and a second minimum was observed (Figure 4, curve 2). This is interpreted as an increase of the film thickness. The best fit of the reflectivity data was obtained with a threelayer model (silicon oxide, OTS and peptide layer). The structural parameters of the hydrophobic OTS were identical to the ones obtained for the bare layer. An additional layer with an electron density of 0.11 e-/Å3 and a thickness of 23 Å was observed. After 72 h incubation time the reflectivity curve showed an increase in the amplitude of the fringe when compared with the profile after 1 h adsorption time (Figure 4, curve 3). In addition, the distance between the two minima changed. These are indications that the thickness and composition of the surface changed. The fitting parameters showed an OTS layer with an electron density of 0.26 e-/Å3 and a thickness of 23 Å. The electron density and thickness of the peptide layer changed to 0.16 e-/Å3 and 20 Å, respectively. The fitting parameters of the reflectivity curves are summarized in Table 2. Figure 5 shows representative topographic images of OTS at different incubation times in B18 solution. The bare OTS layer can be seen in Figure 5a. After 10 min incubation time no significant differences are observed in the layer (Figure 5b) when compared to the bare substrate. In contrast, after 20 min the surface is different from the bare substrate. In this case the surface is covered with amorphous aggregates from which elongated structures arranged in a kind of network are growing (Figure 5c). The surface coverage increased with time and the pattern became more dense (Figure 5d,e). After 72 h adsorption time the surface is covered more uniformly but many small regions without adsorbed molecules were observed (Figure 5f). Section analysis of the images showed that at early adsorption times the heights of the structures were not uniform (Figure 6a). Heights of 25 ( 1 and 37 ( 2 Å were measured after 20 min adsorption time. Similarly, structures of two distinct heights were found after 1 h (23 ( 1 and 32 ( 2 Å) and after 2 h incubation time (20 ( 1 Å; 34 ( 2 Å). Higher structures were preferentially located at the boundary between covered and bare areas and decreased in number with time. After 72 h, the higher areas disappeared and the height amounted to 20 ( 1 Å. The surface roughness after peptide adsorption decreased as a function of time (Figure 6b). Similar trends were found with X-ray reflectivity measurements.
Discussion On a hydrophilic surface, the extent of adsorption is determined by the balance between electrostatic attraction and repulsion in the adsorption layer.24 At pH 7.4, B18 peptide carries a net positive charge. Since the C-terminus of B18 is blocked with an amide group, there are no negatively charged groups in the molecule. (22) Sears, V. F. Neutron News 1992, 3, 26-37. (23) Wasserman, S. R.; Whitesides, G. M.; Tidswell, I. M.; Ocko, B. M.; Pershan, P. S.; Axe, J. D. J. Am. Chem. Soc. 1989, 111, 5852-5861. (24) Haynes, C. A.; Norde, W. Colloids Surf., B 1994, 2, 517-566.
5026 Langmuir, Vol. 23, No. 9, 2007
Rocha et al.
Figure 3. AFM images of silicon oxide layer after incubation in B18 solution for different times. Images were recorded at 1 × 1 µm2 scan. The Z range is 10 nm.
Figure 4. X-ray reflectivity profile of OTS layer after incubation in buffer solution (curve 1) and in B18 solution for 1 h (curve 2) and 72 h (curve 3). The curves were fitted using the parameters of Table 2 (continuous lines in gray). The curves are offset for clarity. Table 2: Structural Parameters Used to Fit the Reflectivity Profiles for the System OTS and B18 Peptidea layer bare layer OTS silicon oxide 1h after adsorption peptide OTS silicon oxide 72h after adsorption peptide OTS silicon oxide
d (Å)
Fe (e-/Å3)
σ (Å)
23.8 ( 0.3 11.0 ( 3.0
0.24 ( 0.01 0.72 ( 0.11
2.7 ( 0.2 7.6 ( 0.2
22.8 ( 0.8 24.0 ( 0.7 11.0 ( 2.0
0.11 ( 0.01 0.24 ( 0.01 0.72 ( 0.01
3.9 ( 0.7 2.6 ( 0.6 7.8 ( 0.3
19.8 ( 0.7 23.1 ( 1.1 11.0 ( 6.0
0.16 ( 0.01 0.26 ( 0.01 0.72 ( 0.02
3.9 ( 0.6 1.0 ( 0.9 7.8 ( 0.6
a d, F and σ are the thickness, scattering length density and roughness, respectively.
Most likely, the absence of B18 adsorption on the positively charged film is due to electrostatic repulsion between the peptide and the surface, whereas adsorption on the negatively charged film is driven by electrostatic attraction. In addition, adsorption may be driven by an entropy gain caused by dehydration of the polyelectrolyte surface and peptide and/or conformational entropy gain resulting from structural changes in the peptide. The adsorbed amount on the negatively charged polymer corresponds to an irreversibly adsorbed state since the surface was rinsed with water after B18 exposure. Adsorption of B18 peptide onto negatively charged surfaces is a slow process. Long incubation times are required to observe irreversible adsorption. One possible reason for the slow adsorption kinetics is that the
peptide aggregates first in solution and it adsorbs in aggregated form. Indeed B18 is known to aggregate in solution at pH >7.4,13 It is not possible to determine unambiguously whether the peptide adsorbs as single molecule or in aggregated form. The most likely mechanism is that the peptide molecules adsorb in nonaggregated state but they require time to conformationally adjust to the surface. Considering that B18 has hydrophobic sequences at both ends of the molecule, it is possible that the peptide molecules undergo conformational changes when interacting with hydrophilic surfaces and once adsorbed, they diffuse on the surface to form clusters. Proteins such as lysozyme are known to diffuse on solid surfaces.25 Additional molecules will preferentially adsorb on the top of those already adsorbed on the surface, building a second layer. The model of a second layer of peptide growing on the clusters is supported by the finding of different aggregate heights by AFM. A possible driving force for adsorption of bulk peptide on already adsorbed molecules may result from the exposed hydrophobic residues. Apparently, this process occurs on a faster time scale than the adsorption of peptide molecules on the bare surface. When bulk peptides come into contact with adsorbed molecules, the hydrophobic residues on each are available for agglomeration, which might be thermodynamically more favorable than adsorption on the bare charged surface. When hydrophobic OTS is exposed to B18 peptide solution, X-ray reflectivity data showed that even after rigorous rinsing with water, there is a layer of peptide adsorbed on the surface. Adsorption of B18 onto hydrophobic OTS is faster than onto hydrophilic surfaces. B18 has a lower activation free energy of adsorption on hydrophobic surfaces than on hydrophilic surfaces. This indicates that B18 shows high affinity for hydrophobic surfaces, which is likely caused by its seven leucine residues, that contribute to a decrease in free energy on going from the aqueous environment (initial state) to the hydrophobic interface (final state). Lower electron density obtained from the X-ray reflectivity data after 1 h adsorption time comparing with 72 h suggests that at early times the coverage of the surface was less uniform. In addition the thickness of the peptide layer was found to decrease with time suggesting diffusion of peptide molecules to the free sites on the substrate, which is supported by the AFM data. The following mechanism for B18 adsorption on OTS layer is proposed. Peptide molecules adsorb irreversibly on the surface after conformational arrangement. Additional molecules will preferentially adsorb from the bulk solution on those already adsorbed. Afterward, molecules will diffuse to the free sites on the substrate. This would explain why the heights of aggregates decrease with time. It is possible that two-dimensional lateral (25) Kim, D. T.; Blanch, H. W.; Radke, C. J. Langmuir 2002, 18, 5841-5850
Fusogenic Peptide B18
Langmuir, Vol. 23, No. 9, 2007 5027
Figure 5. AFM images of OTS layer after incubation in B18 solution for different times. Images were recorded at 1 × 1 µm2 scan. The Z range is 20 nm.
Figure 6. Height (a) and roughness (b) of B18 peptide adsorbed on OTS layer as a function of time as determined by AFM. The square root of time is used to distinguish the experimental points. The mean values and confidence intervals were obtained from at least five independent areas.
diffusion on the surface affects the adsorption process of B18 on OTS layers. It is known that polymers can undergo rapid lateral diffusion on surfaces. The high local concentration of peptide in the amorphous deposit could well act as a reservoir of elongated structures, which would be a more stable state. The growing of amyloid fibrils from amorphous aggregates has been reported for a recombinant amyloidogenic immunoglobulin light chain.26 It is possible that fibril formation by B18 peptide follows a similar mechanism. The presence of amorphous aggregates may serve as nucleation sites for fibril growth. This process might constitute an important mechanism of fibril formation. Although B18 peptide is known to induce fusion of lipid vesicle systems,4,7,13,14 the fusion mechanism is poorly understood. The adsorption behavior of B18 on model solid surfaces may provide information about this process. It is very likely that peptide molecules interact with the membrane surface leading to cluster formation. Hydrophobic interactions will then allow the formation of amorphous aggregates causing a membrane defect. Since B18 has high affinity for hydrophobic surfaces, it will penetrate and (26) Zhu, M.; Souillac, P. O.; Ionescu-Zanetti, C.; Carter, S. A.; Fink, A. L. J. Biol. Chem. 2002, 277, 50914-50922.
diffuse in the target membrane. This result is supported by the finding that B18 inserts into membranes.27 The type of B18 aggregates is though to determine the peptide functional state.4 The fusion activity of B18 is promoted by zinc ions, which induce the formation of globular clusters. In the absence of zinc, B18 aggregates into amyloid fibrils.4 The local concentration is certainly a determinant factor in this process since the peptide forms oligomeric β-sheet structures as well in the presence of zinc and low peptide to lipid ratio. Whether the fibrils are functional relevant for the fusion process is not clear. Selfassociation may contribute to the recruitment of more peptide molecules to the fusion site. However fibril formation seems to result in disruption of the membrane rather than fusion. These might be true as well for amyloidogenic peptides when interacting with cellular membranes. It has been suggested that the lipid membrane might play a catalytic role in the nucleation or selfassembly process of amyloid fibrils, which would lead to a perturbation of the membrane.28 (27) Afonin, S.; Du¨rr, U. H. N.; Glaser, R. W.; Ulrich, A. S. Magn. Reson. Chem. 2004, 42, 195-203. (28) McLaurin, J.; Chakrabartty, A. J. Biol. Chem. 1996, 271, 2648226489.
5028 Langmuir, Vol. 23, No. 9, 2007
Rocha et al.
Figure 7. Schematic representation of B18 adsorption onto solid surfaces. (A) B18 adsorption on negatively charged surfaces. B18 adsorbs onto the negatively charged surface, conformationally adjusts and diffuses on the surface. When a molecule encounters another molecule, they agglomerate, eventually forming clusters (1). Additional molecules will adsorb on the top of those already adsorbed on the surface, building a second layer (2). (B) B18 adsorption on hydrophobic surfaces. Peptide molecules adsorb on the surface after conformational arrangement. Additional molecules will preferentially adsorb from the bulk solution on those already adsorbed and will diffuse to the free sites on the substrate.
Conclusion The degree of adsorption of B18 was determined at pH 7 on different surfaces: positively and negatively charged hydrophilic films, as well as hydrophobic polymers. B18 peptide did not adsorb on hydrophilic positively charged surfaces, which was attributed to electrostatic repulsion between the positively charged peptide and the surface. In contrast, the peptide was found to adsorb on negatively charged surfaces and on hydrophobic surfaces. Figure 7 summarizes the proposed mechanism for the adsorption behavior of B18 on the different surfaces. The adsorption of B18 on negatively charged surfaces was characterized by the formation of globular clusters that are formed due to either diffusion of the molecules on the surface (Figure 7A, pattern 1) or due to growing of a second layer (Figure 7A, pattern 2). The adsorption process of B18 on hydrophobic surfaces is affected by lateral diffusion on the surface (Figure 7B). The peptide showed higher affinity for hydrophobic surfaces than for hydrophilic negatively charged surfaces, which was determined
by adsorption kinetic studies. B18 peptide adsorbed irreversibly on hydrophobic surfaces after 20 min adsorption time, whereas it required more than 1 h to adsorb on negatively charged surfaces. The affinity for hydrophobic surfaces must be caused by its seven leucine residues, which are highly hydrophobic. Acknowledgment. S. Rocha thanks the Fundac¸ a˜o para a Cieˆncia e a Tecnologia (FCT, Portugal) for a Fellowship (BDP/ 25641/2005). These investigations were supported by the German Science Foundation (DFG, Br1378/8), the Max Planck Society, and the FCT research project POCTI/FCB/45365/2002. O. Zscho¨rnig from the University of Leipzig is acknowledged for providing the B18 peptide. We are gratefully acknowledged to Rumen Krastev for assistance during neutron reflectivity measurements and to Anne Heilig for helping during atomic force microscopy measurements. The Hahn-Meitner Institute in Berlin is acknowledged for beam time. LA0628120
