pubs.acs.org/Langmuir © 2010 American Chemical Society
Effect of Surface Charge Distribution on the Adsorption Orientation of Proteins to Lipid Monolayers Sebastian Tiemeyer,* Michael Paulus, and Metin Tolan Fakult€ at Physik/DELTA, Technische Universit€ at Dortmund, Maria-Goeppert-Mayer-Str. 2, D-44221 Dortmund, Germany Received June 29, 2010. Revised Manuscript Received July 26, 2010 The adsorption orientation of the proteins lysozyme and ribonuclease A (RNase A) to a neutral 1,2-dipalmitoyl-snglycero-3-phosphocholine (DPPC) and a negatively charged stearic acid lipid film was investigated by means of X-ray reflectivity. Both proteins adsorbed to the negatively charged lipid monolayer, whereas at the neutral monolayer, no adsorption was observed. For acquiring comprehensive information on the proteins’ adsorption, X-ray reflectivity data were combined with electron densities obtained from crystallographic data. With this method, it is possible to determine the orientation of adsorbed proteins in solution underneath lipid monolayers. While RNase A specifically coupled with its positively charged active site to the negatively charged lipid monolayer, lysozyme prefers an orientation with its long axis parallel to the Langmuir film. In comparison to the electrostatic maps of the proteins, our results can be explained by the discriminative surface charge distribution of lysozyme and RNase A.
Introduction Lipid membranes assign the decisive nature of a biological cell, as they define the interface to the extracellular environment. At this interface, assembled from lipid bilayers, abundant proteinmembrane interactions are taking place contributing to almost all vital processes. For instance, the signal transduction process via the cell membrane is controlled by membrane-associated and soluble proteins.1,2 Furthermore, protein-membrane interactions play a key role in the genesis of many diseases related to protein misfolding, e.g., Alzheimer’s disease or type 2 diabetes mellitus involving islet amyloid polypeptide.3,4 The specific functionality of proteins participating in these processes is mainly given by their complex three-dimensional structure. Thus, studies which yield not only information on protein-membrane interactions5,6 but also on the structural properties of proteins at interfaces are of great interest, but still rare.7,8 Langmuir monolayers formed by lipids are a well-defined system to characterize protein-membrane interactions.9,10 Depending on the headgroup of the lipid, Langmuir monolayers provide different attractive and repulsive interactions between layer and proteins. In this work, the Langmuir monolayers forming the model membranes comprising negatively charged stearic acid lipids and neutral 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) phospholipids, respectively. *To whom correspondence should be addressed. E-mail: sebastian.tiemeyer@ tu-dortmund.de. (1) Cho, W.; Stahelin, R. V. Annu. Rev. Biophys. Biomol. Struct. 2005, 34, 119. (2) Conner, S. D.; Schmid, S. L. Nature 2003, 422, 37. (3) Arispe, N.; Rojas, E.; Pollard, H. B. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 567. (4) Engel, M. F. M.; Khemtemourian, L.; Kleijer, C. C.; Meeldijk, H. J. D.; Jacobs, J.; Verkleij, A. J.; De Kruijff, B.; Killian, J. A.; H€oppener, J. W. M. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 6033. (5) Lipp, M. M.; Lee, K. Y. C.; Zasadzinski, J. A.; Waring, A. J. Science 1996, 273, 1196. (6) Serrano, A. G.; Perez-Gil, J. Chem. Phys. Lipids 2006, 141, 105. (7) Perriman, A. W.; Henderson, M. J.; Evenhuis, C. R.; McGillivray, D. J.; White, J. W. J. Phys. Chem. B 2008, 112, 9532. (8) Chen, C. H.; Malkova, S.; Pingali, S. V.; Long, F.; Garde, S.; Cho, W.; Schlossman, M. L. Biophys. J. 2009, 97, 2794. (9) Brockman, H. Curr. Opin. Struct. Biol. 1999, 9, 438. (10) Kaganer, V. M.; M€ohwald, H.; Dutta, P. Rev. Mod. Phys. 1999, 71, 779.
14064 DOI: 10.1021/la102616h
Since we are interested in the investigation of the influence of a protein’s surface charge distribution on the adsorption behavior to a lipid film, we have chosen the two proteins lysozyme11 and ribonuclease A (RNase A)12 due to their discriminative surface charge distributions. Possessing high isoeletric points (lysozyme pI 11, RNase A pI 9.3), the two proteins carry positive charges at pH 7. While lysozyme exhibits a rather uniform surface charge distribution, RNase A shows a local concentration of positive charges at the active site of the protein. Furthermore, the proteins are globular13 (45 30 30 A˚3 lysozyme, 38 28 22 A˚3 RNase A) and characterized by a remarkably robust conformation. Hence, for the adsorption at a lipid monolayer no significant unfolding is expected.14 In the following, the influence of electrostatic interactions on the adsorption of the proteins is elucidated by the analysis of the structural arrangement of the adsorbed proteins at the monolayerprotein solution interface. Therefore, X-ray reflectivity measurements were conducted which provided electron density profiles perpendicular to the sample surface, whereas calculated electron density profiles were obtained from crystallographic data. The comparison of the experimentally obtained electron density profiles with simulated data yielded the adsorbed proteins’ orientation. The latter was addressed to the surface charge distribution of the proteins by employing electrostatic maps. Moreover, the effect of the protein adsorption on the lateral structure of a stearic acid layer was examined by means of grazing incidence X-ray diffraction.
Experimental Section X-ray Reflectivity and Grazing Incidence Diffraction. X-ray reflectivity (XR) is an interface sensitive technique supplying information about electron density, thickness, and roughness of thin films.15 Under specular condition, (11) Dobson, C. M.; Evans, P. A.; Radford, S. E. Trends Biochem. Sci. 1994, 19, 31. (12) Raines, R. T. Chem. Rev. 1998, 98, 1045. (13) Creighton, T. E. Proteins: Structures and Molecular Properties; W. H. Freeman & Co: New York, 2002. (14) Postel, C.; Abillon, O.; Desbat, B. J. Colloid Interface Sci. 2003, 266, 74. (15) Tolan, M. X-ray Scattering from Soft-Matter Thin Films; Springer: Berlin, 1999.
Published on Web 08/13/2010
Langmuir 2010, 26(17), 14064–14067
Tiemeyer et al.
Article
X-rays with a wavelength λ are reflected at interfaces between media characterized by different electron densities. Since the wave vector transfer in an XR experiment qz = (4π/λ) sin θ possesses only a component perpendicular to the sample’s surface (z-axis), this scattering method is sensitive to the laterally averaged electron density profile along the z direction. In the kinematical approximation,15 the reflectivity is given by Z 2 1 dFðzÞ iqz z e dz Rðqz Þ ¼ RF ðqz Þ Fs dz
)
with the Fresnel reflectivity RF of a sharp interface, the electron density Fs of the substrate, and the electron density profile F(z) perpendicular to the interface. While XR measurements provide only information about vertical electron density profiles, the grazing incidence X-ray diffraction (GIXD) technique resolves the lateral structure of a sample system. A detailed description of this method is given in ref 16. In a GIXD experiment, the incoming X-ray beam hits the sample under a shallow angle which is below the critical angle of total external reflection leading to penetration depths on the order of 5 nm. The scattered radiation is monitored in a horizontal plane parallel to the sample surface. Using this scattering geometry, wave vector transfers q parallel to the surface between 10-3 and several inverse angstroms can be achieved. Thus, this method is well-suited to investigate the lateral structure of lipid monolayers on an aqueous subphase. Sample Preparation and Measurement. The sample material was purchased in the highest available purity from Sigma Aldrich and was used without further purification. DPPC and stearic acid were chosen to prepare a neutral, respectively, negatively charged lipid monolayer. The lipids were dissolved in chloroform to ensure a solution concentration of about 2 mM. All aqueous solutions (pure buffer and protein) were prepared with purified water (resistance of 18.2 MΩ cm) and 10 mM sodium phosphate and adjusted to pH 7.0 using NaOH. Stock protein solutions of bovine RNase A and hen egg white lysozyme with a concentration of 1.0 mg/mL were prepared in buffer solution. A Langmuir trough (KSV Instruments) was used to compress the lipid monolayer with two symmetrical barriers up to a surface pressure of Π = 30 mN/m providing a close-packed condensed Langmuir film. The whole trough was placed in a closed environment in order to prevent evaporation and sample contamination. All reflectivities were recorded with a Bruker AXS D8 advance diffractometer in θ-θ geometry using copper KR radiation (wavelength λ = 1.54 A˚). A G€ obel mirror was applied for beam parallelization leading to a beam size of 0.1 mm in vertical and 10 mm in horizontal direction. The incidence photon flux was 5 107 photons/s. All measurements were performed at room temperature. A typical reflectivity scan covering a qz-range between 0 and 0.4 A˚-1 took about 90 min including the measurement of the diffusely scattered signal. The measurements proceeded in two steps: first, a reference reflectivity of the lipid monolayer on a pure buffer subphase was recorded in order to reveal the electron density profile of the Langmuir film. Afterward, the respective protein was added behind the barriers into the aqueous subphase leading to a concentration of 0.1 mg/mL. Subsequently, the adsorption process was monitored by X-ray reflectivity over a time interval of several hours until an adsorbed protein layer had been constituted. The GIXD measurements of the stearic acid monolayers on an aqueous subphase were performed at the beamline BW1, (16) Jensen, T. R.; Kjaer, K. In Novel Methods to Study Interfacial Layers; M€obius, D., Miller, R., Eds.; Elsevier: Amsterdam, 2001; Vol. 11, p 205. (17) Frahm, R.; Weigelt, J.; Meyer, G.; Materlik, G. Rev. Sci. Instrum. 1995, 66, 1677.
Langmuir 2010, 26(17), 14064–14067
HASYLAB17 using a liquid surface scattering setup. The wavelength of the incoming photons was λ = 1.30 A˚. The stearic acid monolayers were prepared on buffer solution and RNase A solution (concentration c = 0.1 mg/mL) and compressed to a surface pressure of Π = 30 mN/m. GIXD scans were performed in an angular region between θ = 7.25° and θ = 9.25° in order to obtain the size of the crystal domains formed by the lipid chains from the low-order Bragg reflections.
Data Analysis and Simulation For the analysis of the XR measurements, the raw reflectivity data were background-corrected by subtracting the signal of the diffusely scattered radiation. The electron density profile was obtained by fitting the reflectivity data with a least-squares routine employing the Parratt algorithm.18 The effective density model15 was applied to the lipid monolayer assuming a box model consisting of a layer for the head groups and a layer for the aliphatic chains. A sufficient agreement between refinement and data was obtained by dividing the adsorbed protein layer only into three slabs. After refinement, the resulting electron density profile of the adsorbed layer was calculated by subtracting the electron density of the subphase and the lipid monolayer. For the determination of the orientation of the adsorbed proteins, electron density profiles were simulated from crystallographic data taken from the RCSB protein data bank.19 The data sets 1vdq and 1fs3 were used for lysozyme and RNase A, respectively. Every atom in the protein was represented by a sphere with a corresponding van der Waals radius. Assuming a given orientation of the protein with respect to the monolayer, all spheres were projected on an axis perpendicular to the monolayer using a grid of 2 A˚. This procedure provides one-dimensional electron density profiles which depend on the protein’s orientation. Since the data sets did not include hydrogen atoms, the simulated electron density slightly underestimated the measured electron density. For this reason, the simulated electron density was normalized to the maximum value of the data. In addition, electrostatic maps of the proteins were calculated employing the DelPhi software package.20 Results In order to verify that electrostatic interactions are dominant for the observed adsorption process, the interaction of lysozyme and RNase A with the neutral DPPC monolayer was analyzed (see Figure 1). The electron density profiles of the DPPC monolayers are in accordance with previous work.21,22 The two reflectivities taken in the presence of the proteins exhibited no evidence for an adsorbed protein layer. This conclusion is supported by calculations showing that we are able to detect protein layers covering at least 1% of the lipid monolayer. The fact that an adsorbed protein layer could not be observed excludes the existence of a strong attractive interaction between the DPPC monolayer and the two proteins. This can be attributed to the presence of the hydrophilic headgroups preventing an adsorption of proteins due to the hydrophobic interaction which is usually observed at free aqueous solution surfaces.23 (18) Parratt, L. G. Phys. Rev. 1954, 95, 359. (19) RCSB Protein Data Bank 2010; http://www.pdb.org; accessed 05-Jan-2010. (20) Rocchia, W.; Alexov, E.; Honig, B. Phys. Chem. B 2001, 105, 6507. (21) Vaknin, D.; Kjaer, K.; Als-Nielsen, J.; L€osche, M. Biophys. J. 1991, 59, 1325. (22) Wu, G.; Majewski, J.; Ege, C.; Kjaer, K.; Weygand, M. J.; Lee, K. Y. C. L. Biophys. J. 2005, 89, 3159. (23) Yano, Y. F.; Uruga, T.; Tanida, H.; Toyokawa, H.; Terada, Y.; Takagaki, M.; Yamada, H. Langmuir 2009, 25, 32.
DOI: 10.1021/la102616h
14065
Article
Tiemeyer et al.
Figure 1. Left: Reflectivity of the DPPC monolayer before (squares) and after injection of lysozyme (circles) and RNase A (triangles) into the subphase, respectively. Reflectivities are shifted by an order of magnitude. Refined reflectivities are shown as solid lines. Right: Electron density profiles obtained by the refinement. The black line represents the box model of the DPPC monolayer. All profiles are normalized to the aqueous subphase.
Figure 2. Left: Reflectivity of the stearic acid monolayer before (squares) and after (circles) injection of lysozyme into the subphase. Reflectivities are shifted by an order of magnitude. Refined reflectivities are shown as solid lines. Right: Electron density profiles obtained by the refinement. The black line represents the box model of the stearic acid monolayer. All profiles are normalized to the aqueous subphase.
Figure 2 shows reflectivity data and electron density profiles for a stearic acid monolayer on an aqueous subphase without (squares) and with (circles) the presence of lysozyme. Before injection, the reflectivity of the stearic acid monolayer exhibits a single oscillation (Kiessig fringe), whereas after injection of lysozyme, the reflectivity is characterized by a superposition of several oscillations caused by an adsorbed protein layer beneath the lipid monolayer. On the right-hand side of Figure 2, the corresponding electron density profiles are depicted. The electron density in the regime of the lipid monolayer is obviously unaffected by the adsorbed layer. Hence, the XR measurements denote that the lysozyme molecules adsorbed without disturbing the structural integrity of the overlying lipid monolayer. The thickness of the adsorbed protein layer was dLys = 30 ( 0.5 A˚, which indicated a parallel orientation of the long axis of lysozyme relative to the lipid monolayer (these results are supported by Postel et al.14). Compared to the electron density of lysozyme calculated from its number of electrons and its partial specific volume (PSV),24 the electron density of the adsorbed lysozyme layer conforms to a closed protein monolayer. In order to determine more precisely the orientation of the lysozyme molecules, electron density profiles along the main axes of the protein were calculated. Figure 3 illustrates on the left-hand side the protein orientation beneath the stearic acid monolayer, whereas the right-hand side (24) Harpaz, Y.; Gerstein, M.; Chothia, C. Structure 1994, 2, 641.
14066 DOI: 10.1021/la102616h
Figure 3. Left: Assumed orientation of lysozyme with electrostatic maps under a negatively charged stearic acid monolayer. Positively and negatively charged regions are denoted blue and red, respectively. Right: Comparisons between the electron density profile obtained from the reflectivity and simulated profiles from crystallographic data taking into account the protein orientation shown on the left-hand side.
shows the experimental electron density profile compared to the respective simulated density profiles. Apparently, an orientation along the z-axis can be excluded due to an enormous mismatch between the measured and simulated electron density profiles. The similarity of the simulated electron density profiles along x- and y-axes with the profile based on the measurement indicates a superposition of the protein orientation along the two axes. A superposition of the two different orientations is strongly suggested by the homogeneous distribution of charge on the surface of lysozyme (see Figure 3). As a consequence, the protein does not prefer a special orientation if the condition of close contact to the negatively charged layer is fulfilled. Thus, only the orientation with the long axis perpendicular to the layer, which can be excluded due to the experimental data, is energetically inappropriate. RNase A, which has in contrast to lysozyme a nonuniform surface charge distribution, exhibits a different adsorption behavior. In Figure 4, reflectivity data and electron density profiles are shown for the adsorption of RNase A on a stearic acid monolayer. On the left-hand side, the reflectivity of the stearic acid monolayer is displayed before injection of RNase A into the subphase (squares) and after the injection (circles). On the right-hand side of Figure 4, the electron density profile shows an adsorbed RNase A layer after injection of the protein. As for lysozyme, the electron density in the regime of the lipid monolayer was unaffected by the adsorbed layer. The thickness of the protein layer was dRNase = 32 ( 0.5 A˚. Compared to the tabulated dimensions of RNase A (38 28 22 A˚3), the measured layer thickness suggests an orientation of the 28 A˚ long axis of RNase A perpendicular Langmuir 2010, 26(17), 14064–14067
Tiemeyer et al.
Figure 4. Left: Reference reflectivity of the stearic acid monolayer (squares) and a reflectivity (circles) after injection of RNase A into the subphase. Reflectivities are normalized to the Fresnel reflectivity and shifted by an order of magnitude, respectively. Refined reflectivities are shown as solid lines. Right: Electron density profiles obtained from the reflectivities. The black line represents the box model of the stearic acid monolayer. The profiles are normalized to the electron density of the aqueous subphase.
Article
Figure 6. GIXD data of a stearic acid monolayer on a pure buffer (square) and on a RNase A containing subphase (circles). Solid lines denote Loretzian fits to the data.
with the data. However, the slightly broadened experimental profile might be caused by thermal fluctuations of the interface or due to slight disorder within the protein layer. In the y orientation, the active site of RNase A is in opposition to the negatively charged lipid monolayer. Considering the electrostatic map of RNase A, this result is supported by the fact that the main part of the positive surface charges is concentrated at the active site. Moreover, rotating the protein by an angle of (15° on an axis parallel to the lipid film significantly degrades the agreement of simulated and experimental density profiles. Thus, RNase A couples in a very specific manner to the lipid monolayer. For the purpose of verifying that the influence of the protein adsorption did not affected the lateral structure of the stearic acid film, GIXD measurements were conducted (see Figure 6). Therefore, the size of the crystalline domains of the lipid monolayer was determined by the width of the recorded Bragg reflections. The shape of the Bragg reflections was refined by a Lorentzian yielding the full width at half-maximum (FWHM). Using the Scherrer formula,16 the domain sizes were calculated to Lbuffer = 247 ( 25 A˚ and LRNase = 261 ( 25 A˚. The difference between these two values is within the experimental error. Hence, the adsorbed RNase A layer did not affect the lipid monolayer.
Summary
Figure 5. Left: Assumed orientation of RNase A with electrostatic maps under a negatively charged stearic acid monolayer. Positively and negatively charged regions are denoted blue and red, respectively. Right: Comparisons between the electron density profile obtained from the reflectivity and simulated profiles calculated from crystallographic data taking into account the protein orientation shown on the left-hand side.
to the lipid layer. As in the case of lysozyme, the electron density of RNase A calculated from the protein’s PSV24 matched the experimental electron density. Accordingly, full protein monolayer coverage of the stearic acid Langmuir film was observed. The comparison of simulated and experimental electron density profiles are shown in Figure 5. Assuming a protein orientation along the x- and z-axes, respectively, the simulated and measured electron density profiles do not match. In contrast, the simulated electron density profile along the y-axis shows excellent agreement Langmuir 2010, 26(17), 14064–14067
In conclusion, we have revealed the adsorption orientation for the proteins lysozyme and RNase A underneath negatively charged stearic acid monolayers. Moreover, we could address the respective adsorption orientation to the surface charge distribution of the proteins. In the case of RNase A, the comparison of experimental and simulated electron density profiles has elucidated the most accurate orientation of the protein pointing its positively charged active site toward the negative lipid monolayer. Due to the homogeneous surface charge distribution of lysozyme, for this protein the orientation of the long axis perpendicular to the water-lipid monolayer interface was strongly suppressed, whereas no specific adsorption orientation with respect to the lipid monolayer has emerged. To conclude in general, we have shown that our approach of combining experimental electron density profiles with simulated ones from crystallographic data is a very useful tool to obtain new information about adsorption processes in biological systems. Acknowledgment. S. Tiemeyer thanks the NRW Forschungsschule ’Forschung mit Synchrotronstrahlung in den Nano- und Biowissenschaften’ for financial support. The authors acknowledge B. Struth for technical support and HASYLAB (DESY Hamburg) for providing synchrotron radiation. DOI: 10.1021/la102616h
14067
