Protein-Directed Assembly of Binary Monolayers at the Interface and

Engineering, Nanjing UniVersity, Nanjing 210093, P. R. China, and School of Chemistry and Chemical. Engineering, Yangzhou UniVersity, Yangzhou 225002,...
3 downloads 0 Views 440KB Size
Download PDF
8142

Langmuir 2007, 23, 8142-8149

Protein-Directed Assembly of Binary Monolayers at the Interface and Surface Patterns of Protein on the Monolayers Xuezhong Du,*,† Yuchun Wang,† Yuanhua Ding,‡ and Rong Guo*,‡ Key Laboratory of Mesoscopic Chemistry (Ministry of Education), School of Chemistry and Chemical Engineering, Nanjing UniVersity, Nanjing 210093, P. R. China, and School of Chemistry and Chemical Engineering, Yangzhou UniVersity, Yangzhou 225002, P. R. China ReceiVed April 3, 2007. In Final Form: May 20, 2007 Ferritin-directed assembly of binary monolayers of zwitterionic dipalmitoylphosphatidylcholine and cationic dioctadecyldimethylammonium bromide (DOMA) at the interface and surface patterns of ferritin on the monolayers have been investigated using a combination of infrared reflection absorption spectroscopy, surface plasmon resonance, and atomic force microscopy. Ferritin binding to the binary monolayers at the air-water interface at the surface pressure 30 mN/m, primarily driven by the electrostatic interaction, gives rise to a change in tilt angle of hydrocarbon chains from 15° ( 1° to 10° ( 1° with respect to the normal of the monolayer at the mole fraction of DOMA (XDOMA) of 0.1. The chains at XDOMA ) 0.3 are oriented vertical to the water surface before and after protein binding. A new mechanism for protein binding to the binary monolayers is proposed. The secondary structures of the adsorbed ferritin are prevented from changing to some extent due to the existence of the monolayers. The amounts of the bound protein on the monolayers at the air-water interface are increased in comparison with those on the preimmobilized monolayers at low XDOMA. The increased amounts and different patterns of the adsorbed protein at the monolayers are mostly attributed to the formation of multiple binding sites available for ferritin, which is due to the lateral reorganization of the lipid components in the monolayers induced by the protein in the subphase. The created multiple binding sites on the monolayer surfaces through the protein-directed assembly can be preserved for subsequent protein binding.

Introduction In a variety of biological processes, protein-membrane interactions are often multivalent, and their properties can differ markedly from the corresponding monovalent constituents.1 Multivalent interactions are crucial for discovering a broad range of biosensors with various functions for use in medicine, environment, fermentation, and food processing. Langmuir monolayers have half a framework structure of cellular membranes and can mimic cell surface rearrangement due to the lateral mobility of the monolayers at the air-water interface. Lateral mobility is expected to play a key role in free rearrangement of ligand molecules in membranes,2-4 because a ligand for the second and subsequent binding events can be delivered through the lateral rearrangement of monolayer components.5,6 The multivalent protein binding can be realized through the interactions with multicomponent unilamellar vesicles or fluid-supported lipid bilayers (SLBs) rearranged from the unilamellar vesicles,5,7 but the mole fractions of functional lipids on the targeted surfaces cannot be precisely controlled, because the phospholipids with different headgroups are heterogeneously distributed on the two leaflets of vesicles. In comparison with * To whom correspondence should be addressed. Fax: 86-25-83317761. E-mail: [email protected] (X.D.); [email protected] (R.G.). † Nanjing University. ‡ Yangzhou University. (1) Kiessling, L. L.; Gestwicki, J. E.; Strong, L. E. Angew. Chem. Int. Ed. 2006, 45, 2348. (2) Groves, J. T.; Ulman, N.; Boxer, S. G. Science 1997, 275, 651. (3) Spencelayh, M. J.; Cheng, Y.; Bushby, R. J.; Bugg, T. D. H.; Li, J.-J.; Henderson, P. J. F.; O’ Reilly, J.; Evans, S. D. Angew. Chem. Int. Ed. 2006, 45, 2111. (4) Philips, K. S.; Wilkop, T.; Wu, J.-J.; Al-Kaysi, R. O.; Cheng, Q. J. Am. Chem. Soc. 2006, 128, 9590. (5) Yang, T.; Baryshnikova, O. K.; Mao, H.; Holden, M. A.; Cremer, P. S. J. Am. Chem. Soc. 2003, 125, 4779. (6) Du, X.; Hlady, V.; Britt, D. Biosens. Bioelectron. 2005, 20, 2053. (7) Bondurant, B.; Last, J. A.; Waggoner, T. A.; Slade, A.; Sasaki, D. Y. Langmuir 2003, 19, 1829.

the fluid-SLBs (a thin layer water of about 1 nm thick between the bilayer and an underlying support8), the molar ratios of different components in the Langmuir monolayers can be precisely controlled. Besides, the lipid components in the monolayers are capable of lateral mobility. A number of studies on protein adsorption at the self-assembled monolayers (SAMs) consisting of binary mixtures have been reported,9-11 but the multivalent interactions between the proteins and functional lipids are obviously depressed, probably due to the limitation of lateral mobility through the immobilization of covalent attachment. On the other hand, the molar ratios of different components in the SAMs cannot be precisely controlled, because they are generally different from the original molar ratios in the solutions from which the SAMs are formed.12 Patterned surfaces based on the SAMs can be generated using a few techniques, such as microcontact printing,13,14 photolithography,15 and dip-pen nanolithography.16 Protein patterns are readily formed through the apparently “selective” adsorption of proteins in the regions of functional lipids. However, the formation of the protein patterns is determined by the preparation of the patterned monolayer surfaces; in addition, the protein adsorption in the regions is mostly nonselective. This study is motivated by protein-directed binding to the fluid monolayers to generate tailor-made surfaces with a high protein affinity for biosensing and information storage. (8) Johnson, S. J.; Bayerl, T. M.; McDermott, D. C.; Adam, G. W.; Rennie, A. R.; Thomas, R. K.; Sackmann, E. Biophys. J. 1991, 59, 289. (9) Prime, K. L.; Whitesides, G. M. Science 1991, 252, 1164. (10) Gupta, V. K.; Abbott, N. L. Science 1997, 276, 1533. (11) Hodneland, C. D.; Lee, Y.-S.; Min, D.-H.; Mrksich, M. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 5048. (12) Bain, C. D.; Whitesides, G. M. J. Am. Chem. Soc. 1988, 110, 3665. (13) Xia, Y.; Whitesides, G. M. Angew. Chem. Int. Ed. 1998, 37, 550. (14) Tinazli, A.; Tang, J.; Valiokas, R.; Picuric, S.; Lata, S.; Piehler, J.; Liedberg, B.; Tampe´, R. Chem. Eur. J. 2005, 11, 5249. (15) Folch, A.; Toner, M. Annu. ReV. Biomed. Eng. 2000, 2, 227. (16) Demers, L. M.; Ginger, D. S.; Park, S.-J.; Li, Z.; Chung, S.-W.; Mirkin, C. A. Science 2002, 296, 1836.

10.1021/la700955f CCC: $37.00 © 2007 American Chemical Society Published on Web 06/21/2007

Protein-Directed Assembly of Binary Monolayers

In this paper, the zwitterionic phospholipids dipalmitoylphosphatidylcholine (DPPC) are used to constitute a protein-resistant matrix monolayer in which cationic surfactants dioctadecyldimethylammonium bromide (DOMA) are distributed. The binary monolayers at the air-water interface are miscible and stable due to the attractive electrostatic interaction between their headgroups.17,18 Ferritin (pI 4.5) binding then proceeds to achieve the multiple interactions before the binary monolayers at the air-water interface are immobilized using the horizontal Langmuir-Blodgett (LB) technique. Ferritin is an almost spherical protein with a molecular weight of 680 kDa and a diameter of 12.5 nm. Ferritin has an isoelectric point (pI) of 4.5, and its net charge is negative at pH > 5. Ferritin is selected as a model protein, due to its stable, compact, and nearly spherical structure, which could eliminate discrepancy in adsorption orientation and structure deformation upon binding. The surfaces with a high protein affinity are created through the directed assembly of the binary monolayers induced by the protein in the subphase. Lots of techniques have been used to study the interaction of proteins with monolayers. Infrared reflection absorption spectroscopy (IRRAS) is one of the leading methods for structural analyses of monolayers at the air-water interface.19 This technique is currently the only physical method able to directly monitor protein secondary structure in situ in Langmuir monolayers.20 Surface plasmon resonance (SPR) is an excellent method for the investigation of protein adsorption (at the solid-water interface) in real time without labeling analytes and gives rich information about protein binding kinetics and surface density. Atomic force microscopy (AFM) is an important tool for directly visualizing topographic surfaces on a nanometer scale and can observe the spatial distribution of adsorbed proteins. One of its most significant advantages is the ability to image the surfaces in solution, so that proteins are maintained in an aqueous environment. The IRRAS, SPR, and AFM techniques are suited for the studies of interfacial systems relevant to aqueous solutions, such as the air-water interface and the solid-water interface. The physiological environments of proteins can be maintained for the achievement of protein recognition and biosensing. Moreover, abundant information on the interaction of proteins with the monolayers at the interface can be obtained on the molecular level, in real time, and at spatial distribution using a combination of the three techniques. Experimental Section Chemicals. Synthetic L-R-DPPC (∼99%) and DOMA (>99%) were purchased from Sigma and Acros Organics, respectively, and used as received. Their stock solutions were prepared in pretreated chloroform (analytical grade) at a concentration of 1 mM and stored at -20 °C prior to use. The mixtures of DPPC and DOMA were prepared volumetrically from their stock solutions. 1-Octadecanethiol (ODT, 95%) and octadecyltrichlorosilane (OTS) were purchased from Fluka and Aldrich, respectively. Triton X-100, ethanol, NaCl, and NaOH were of analytical grade. Water used was double-distilled (pH 5.6, resistivity 18.2 MΩ cm) after a deionized exchange. Horse spleen ferritin (76 mg/mL, type 1) in 0.15 M NaCl was supplied by Sigma and used as received. Ferritin solution was prepared by diluting the stock solution with double-distilled water to desired concentrations. IRRAS Measurements. IRRAS spectra of the monolayers at the air-water interface were recorded on a Bruker Equinox 55 FTIR (17) Gonc¸ alves da Silva, A. M.; Roma˜o, R. I. S. Chem. Phys. Lipids 2005, 137, 62. (18) Wang, Y.; Du, X. Langmuir 2006, 22, 6195. (19) Mendelsohn, R.; Brauner, J. W.; Gericke, A. Annu. ReV. Phys. Chem. 1995, 46, 305. (20) Cai, P.; Flach, C. R.; Mendelsohn, R. Biochemistry 2003, 42, 9446.

Langmuir, Vol. 23, No. 15, 2007 8143 spectrometer connected to an XA-511 external reflection attachment (Bruker, Germany) with a shuttle trough system and a liquid-nitrogencooled MCT detector. Sample (film-covered surface) and reference (film-free surface) troughs were fixed on a shuttle device driven by a computer-controlled stepper motor for allowing spectral collections from the two troughs in an alternating fashion. A KRS-5 polarizer was used to generate perpendicularly polarized lights, and the efficiency of the polarizer was determined to be about 99.2%.21 The IRRAS experiments were carried out at 22 °C. The film-forming molecules were spread from chloroform solution of desired volumes, and 20 min was allowed for solvent evaporation. The measurement system was then enclosed for humidity equilibrium and monolayer relaxation for 4 h prior to compression. The monolayers were compressed discontinuously to the desired surface pressure of 30 mN/m from ∼0 mN/m. After 30 min relaxation, the two moving barriers were stopped and the monolayer areas were kept constant. For the experiments on protein binding, concentrated ferritin solutions were injected into the unstirred subphase beneath the compressed monolayers to achieve a final protein concentration of 1.0 µg/mL. The external reflection absorption spectrum of pure water was used as a reference. The spectra were recorded with a resolution of 8 cm-1 by coaddition of 1024 scans. A time delay of 30 s was allowed for film equilibrium between trough movement and data collection. Spectra were acquired using p-polarized radiation followed by s-polarized light. During the data collection at different angles of incidence, surface pressure changed slightly for the monolayers (e2 mN/m). The IRRAS spectra were used without any processing or baseline correction. The attenuated total reflection FTIR (ATRFTIR) spectrum of aqueous ferritin solution (10 mg/mL) was measured on the Equinox 55 FTIR spectrometer with a DTGS detector at the resolution 8 cm-1. The ATR accessory contained a ZnSe crystal at a nominal angle of incidence 45°. The final spectrum of ferritin was obtained after the spectrum of water was subtracted. SPR Measurements. A Teflon trough with a dimension of 4 cm × 2 cm × 1 cm was home-built similar to the microtrough used before.6,18 The trough walls were undercut by 45° to eliminate the formation of a meniscus presenting a planar interface,6 on which an integrated optics SPR sensor (Spreeta, Texas Instruments) was used to measure protein adsorption kinetics for each monolayer.6,18 The Spreeta sensors combine the sensor surfaces with all optic and electronic components required for SPR experiments in a compact and lightweight assembly.22,23 The SPR sensor was first cleaned using an aqueous solution of 1% Triton X-100 and 0.1 M NaOH followed by double-distilled water. Its sensing gold surface was made hydrophobic by immersion in a 2 mM ODT solution in absolute ethanol for 20 min followed by rinsing with double-distilled water. The SPR sensor was then dried and positioned above the monolayer at the air-water interface. The SPR sensor was initialized in air and calibrated in double-distilled water, and a SPR baseline was obtained in water. A binary monolayer of DPPC and DOMA was spread until the desired surface pressure 30 mN/m was reached, and then it was allowed for relaxation for 1 h. In the case of the control experiment (a monolayer at the solid-water interface), the ODT-modified SPR sensor was slowly lowered into contact with the monolayer and brought to a depth of ca. 2 mm using a micromanipulator. This procedure was equivalent to a horizontal monolayer transfer to a solid substrate but without ever removing the substrate (i.e., the sensor) carrying the transferred monolayer from the trough.6 Upon contact of the SPR sensor with the monolayer, a step increase of the SPR signal was recorded and a new SPR baseline was established for a period of 10 min to ensure the integrity of the transferred monolayer prior to injecting protein or exchanging the subphase. The protein solution of desired volumes was injected into the subphase to reach a final ferritin concentration of 4.0 µg/mL. The protein binding was allowed to proceed for at least 2 h, the subphase was then exchanged with acidic water of pH 3.0 to remove protein from the monolayer and finally with pure water of pH 5.6 prior to (21) Du, X.; Wang, Y. J. Phys. Chem. B 2007, 111, 2347. (22) Weimar, T. Angew. Chem. Int. Ed. 2000, 39, 1219. (23) Whelan, R. J.; Wohland, T.; Neumann, L.; Huang, B.; Kobilka, B. K.; Zare, R. N. Anal. Chem. 2002, 74, 4570.

8144 Langmuir, Vol. 23, No. 15, 2007

Du et al.

Figure 1. Time-dependent s-polarized IRRAS spectra of a DPPC monolayer at the surface pressure of 30 mN/m at the angle of incidence of 30° upon injecting ferritin with a final concentration of 1.0 µg/ mL. reintroducing protein for binding. In the case of the fluid monolayer (a monolayer at the air-water interface), protein was injected beneath the compressed monolayer after 1 h relaxation and allowed for binding at least for 2 h. An ODT-modified SPR sensor was brought into contact with the protein-bound monolayer, and SPR signals were recorded. The following procedures were the same as those in the case of the control experiment. For the initial protein adsorption on the fluid monolayers, the binding kinetics were not attained, as the SPR sensors could not be in contact with the monolayers if they were to remain fluid. However, the final adsorption values were obtained by placing the SPR sensors in contact with the monolayers after 2 h of binding. At a given molar ratio, the two types of monolayers nearly had the same molecular densities because the identical amounts of the samples were spread at the fixed trough area. This meant that the SPR angle shifts, contributing from the binary lipid and ODT monolayers on the SPR sensors, were practically identical in the two cases. The SPR angle shifts due to the protein adsorption were only presented after the subtraction of the contributions from the lipid and ODT monolayers. Monolayer Transfer and AFM Measurements. The same microtrough and setup were used for the transfer of ferritin-bound monolayers as in the subsection “SPR Measurements” under the identical conditions. The OTS-modified silicon wafers were adhered to the gold surfaces of SPR sensors for the immobilization of monolayers before and after ferritin binding by the LB technique: the OTS-modified silicon wafer was horizontally pressed into the subphase through the interface covered with the monolayer. The remaining lipids were removed from the interface. The wafer was kept in water during the period of further processing and AFM imaging. AFM images were acquired on an AFM Nanoscope III (Veeco, Digital Instruments) with the contact mode in air or aqueous solution. The cantilever used was commercially available from Park Scientific Instruments, with a normal force constant of 0.01 N m-1 used. The probe was irradiated under a high-intensity UV lamp for about 15 min just prior to imaging to remove organic contaminants from the tip surface.

Results and Discussion Protein Binding to the Monolayers at the Air-Water Interface. Figure 1 shows time-dependent IRRAS spectra of a DPPC monolayer at the surface pressure 30 mN/m after the injection of ferritin for s-polarization (polarized perpendicular to the plane of incidence). The spectrum after 6 h is practically identical to that before protein injection, which indicates that the ferritin does not bind to the condensed zwitterionic monolayer.

Figure 2. Time-dependent s-polarized IRRAS spectra of binary monolayers of DPPC and DOMA at the surface pressure of 30 mN/m at the angle of incidence of 30° upon injecting ferritin with a final concentration of 1.0 µg/mL: (a) XDOMA ) 0.1, (b) XDOMA ) 0.3, (c) XDOMA ) 0.3.

The zwitterionic headgroups are known to be protein-resistant,24,25 and no protein adsorption at the zwitterionic monolayers is observed at the surface pressures above 30 mN/m.26-30 The binary monolayers of DPPC and DOMA are miscible and stable and show a condensation feature in comparison with the monolayers (24) Murphy, E. F.; Lu, J. R.; Brewer, J.; Ressell, J.; Penfold, J. Langmuir 1999, 15, 1313. (25) Chen, S.; Zheng, J.; Li, L.; Jiang, S. J. Am. Chem. Soc. 2005, 127, 14473. (26) Bi, X.; Taneva, S.; Keough, K. M. W.; Mendelsohn, R.; Flach, C. R. Biochemistry 2001, 40, 13659. (27) Maltseva, E.; Kerth, A.; Blume, A.; Mo¨hwald, H.; Brezesinski, S. ChemBioChem 2005, 6, 1817. (28) Maltseva, E.; Brezesinski, G. ChemPhysChem 2004, 5, 1185. (29) Kim, S. H.; Franses, E. I. J. Colloid Interface Sci. 2006, 295, 84. (30) Phang, T.-L.; McClellan, S. J.; Franses, E. I. Langmuir 2005, 21, 10140.

Protein-Directed Assembly of Binary Monolayers

Langmuir, Vol. 23, No. 15, 2007 8145 Table 1. Parameters Used in IRRAS Band Simulations and Tilt Angles of Hydrocarbon Chainsa system XDOMA ) 0.1 (water) XDOMA ) 0.1 (ferritin) XDOMA ) 0.3 (water) XDOMA ) 0.3 (ferritin)

θ R (deg) kmax (deg) L (Å) nord 15 10 0 0

0.60 0.60 0.65 0.65

90 90 90 90

2.05 2.05 2.05 2.05

1.46 1.46 1.46 1.46

next Γ (%) 1.57 1.57 1.57 1.57

99.2 99.2 99.2 99.2

a θ, tilt angle of chain axis; kmax, maximum extinction coefficient; R, angle between νa(CH2) vibrational dipole moment and chain axis; L, extented chain length (taking into account monolayer thickness); nord and next, ordinary and extraordinary refractive indexes;19 Γ, polarizer efficiency.21

Figure 3. Difference IRRAS spectra before and after ferritin adsorption at XDOMA ) 0.1 and 0.3 in comparison with IRRAS spectrum of ferritin at the air-water interface free of lipid monolayer and ATR-FTIR spectrum of aqueous ferritin solution (10 mg/mL).

of the two individual components due to the electrostatic interaction that occurs between the zwitterionic and cationic lipid headgroups.17,18 For the binary monolayers at the mole fractions of DOMA (XDOMA) of 0.1 and 0.3 shown in Figure 2a,b, ferritin injection results in the appearance of amide I and II bands and an increase in the intensity of water bands (Figure 2c). The latter indicates an increasing surface layer thickness due to protein adsorption.27 Blume and co-workers investigated the relation between the intensity of the ν(OH) band at approximately 3600 cm-1 and the film thickness at the air-water interface.31 They could correlate the increase in the intensity of the band directly with the increase of film thickness.31 The new bands at 1657 and 1550 cm-1 are due to the amide I (primarily from the peptide bond CdO stretching mode) and amide II bands (from C-N stretching and N-H bending vibrations) of ferritin, respectively. From the difference spectra after and before protein adsorption, the two bands are clearly observed. This indicates that the protein binding is primarily driven by the electrostatic interaction (net negative charges for ferritin at pH 5.6). Ferritin hardly penetrates into the monolayers at 30 mN/m18 but bind to the hydrophilic headgroups. Figure 3 shows IRRAS spectra of adsorbed ferritin at the binary monolayers at the air-water interface in comparison with ATR-FTIR spectrum of aqueous ferritin solution and IRRAS spectrum of ferritin film at the air-water interface (free of monolayer). Previous studies have been shown that a direct correlation exists between the frequency of the amide I band and protein secondary structure.32,33 In comparison with an aqueous ferritin solution, the amide I band of the adsorbed ferritin at the monolayers increase in frequency by about 5 cm-1, suggesting a change in conformation of ferritin upon binding to the monolayers. However, the magnitude of frequency shift in the amide I band is relatively small in comparison with the amide I band (1659 cm-1) of ferritin directly adsorbed at the air-water interface free of lipid monolayer. This indicates that the hydrophilic headgroups of the binary monolayers prevent protein from changing conformation to some extent. For s-polarized radiation, the bands are always negative and their intensities decrease with increasing angle of incidence. For p-polarized radiation, which is polarized parallel to the plane of incidence, the bands are initially negative and their intensities (31) Hussain, H.; Kerth, A.; Blume, A.; Kressler, J. J. Phys. Chem. B 2004, 108, 9962. (32) Krimm, S.; Bandekar, J. AdV. Protein Chem. 1986, 38, 181. (33) Dousseau, F.; Pezolet, M. Biochemistry 1990, 29, 8771.

increase with increasing angle of incidence and reach a maximum, and then a minimum in the reflectivity is found at the Brewster angle. Beyond the Brewster angle, the bands become positive and their intensities decrease upon further increase of incidence angle. It is known that the IRRAS data are defined as plots of reflectance-absorbance (RA) versus wavenumber. RA is defined as -log(R/R0), where R and R0 are the reflectivities of the filmcovered and film-free surfaces, respectively. Herein, the Kuzmin and Michailov’s model34,35 developed by Gericke et al.19,36,37 is used to describe IRRAS band intensities. The related parameters for the simulation of chain tilt angles are listed in Table 1. The parameter kmax is related to the extinction coefficient for the vibration mode and the density of the film-forming molecules at the air-water interface.27 The relative magnitude of the molecular density can be obtained from the appropriate surface pressure-area isotherms. The protein binding is carried out after the monolayers are compressed to about 30 mN/m for relaxation at the fixed trough area; thus, the number of lipid molecules on the surface does not vary and the frequencies of the vibrational modes do not change, i.e., the chain conformation is not altered, as the protein is introduced. The same kmax values are used for fitting the experimental data before and after protein injection.21 Chain tilt angles θ are determined by measuring νa(CH2) band intensities with p- and s-polarized lights as a function of incidence angle (the corresponding IRRAS spectra shown in Supporting Information) and using an established procedure37 for the simulation. Figure 4 shows the theoretical RA values (lines) and experimental data (symbols) of the νa(CH2) bands for the binary monolayers after ferritin adsorption saturation against angles of incidence. At XDOMA ) 0.1, the tilt angle of the hydrocarbon chains is a little decreased to 10° ( 1° after ferritin binding from 15° ( 1° prior to protein binding; however, the chains increase in tilt angle up to 25°-30° upon human serum albumin (HSA) binding.27 At XDOMA ) 0.3, the chains remain almost vertical to the water surface before and after ferritin binding (a tilt angle of 0° ( 3°), which is consistent with the behavior of the most condensed monolayer in this case.18 Figure 5 shows the intensity ratio of the νa(CH2) band to the νs(CH2) one [νa(CH2)/νs(CH2)] at XDOMA ) 0.1 and 0.3 before and after ferritin adsorption for p- and s-polarization against angle of incidence, respectively. The intensity ratio after ferritin adsorption undergoes a small increase or decrease in comparison with that before protein adsorption for the respective monolayers, which is most likely due to the occurrence of chain twist. The intensity ratio of νa(CH2)/νs(CH2) reflects the twisted orientation (34) Kuzmin, V. L.; Michailov, A. V. Opt. Spectrosc. 1981, 51, 383. (35) Kuzmin, V. L.; Romanov, V. P.; Michailov, A. V. Opt. Spectrosc. 1992, 73, 1. (36) Gericke, A.; Michailov, A. V.; Hu¨hnerfuss, H. Vib. Spectrosc. 1993, 4, 335. (37) Flach, C. R.; Gericke, A.; Mendelsohn, R. J. Phys. Chem. B 1997, 101, 58.

8146 Langmuir, Vol. 23, No. 15, 2007

Figure 4. Comparison of the simulated (lines) and measured (symbols) RA values of νa(CH2) bands for the monolayers at the surface pressure of 30 mN/m after ferritin saturation adsorption for p- and s-polarization: (a) θ ) 10° ( 1° and kmax ) 0.60 at XDOMA ) 0.1, (b) θ ) 0° ( 3° and kmax ) 0.65 at XDOMA ) 0.3.

of hydrocarbon chains.38-41 It is known that both CH2 transition moment directions are perpendicular to the hydrocarbon chains, with the νs(CH2) transition moment being oriented along the bisector of the H-C-H bond angle, while the νa(CH2) transition moment is perpendicular to this. After protein binding, the axes of the hydrocarbon chains at XDOMA ) 0.1 and 0.3 are oriented almost vertical to the water surface at the tilt angles 0°-10°. For s-polarization of the IR beam, stretching band intensities will become maximal when all transition moments are oriented horizontally. However, the changes in intensity ratio of νa(CH2)/ νs(CH2) indicate that the C-C-C planes of the chains at XDOMA ) 0.1 are slightly parallel to the barriers of the Langmuir trough, and the C-C-C planes at XDOMA ) 0.3 are somewhat perpendicular to the barriers. The orientation of the PC headgroup is generally referred to as the tilt angle of the P-N (phosphorus and nitrogen) dipole moment with respect to the surface.42 Numerous studies have been completed on phospholipid bilayer systems, and it has been suggested that the P-N dipole almost lies flat with respect to the membrane surface.43-45 In the DPPC monolayer, the sizes (38) Ren, Y.; Kato, T. Langmuir 2002, 18, 6699. (39) Du, X.; Miao, W.; Liang, Y. J. Phys. Chem. B 2005, 109, 7428. (40) Wang, Y.; Du, X.; Guo, L.; Liu, H. J. Chem. Phys. 2006, 124, 134706. (41) Wang, Y.; Du, X.; Miao, W.; Liang, Y. J. Phys. Chem. B 2006, 110, 4914. (42) Hauser, H.; Pascher, I.; Pearson, R. H.; Sundell, S. Biochim. Biophys. Acta 1981, 650, 21. (43) Ma, G.; Allen, H. C. Langmuir 2006, 22, 5341. (44) Gennis, R. B. Biomembranes: Molecular Structure and Functions; Springer-Verlag: New York, 1989. (45) Dominguez, H.; Smondyrev, A. M.; Berkowitz, M. L. J. Phys. Chem. B 1999, 103, 9582.

Du et al.

Figure 5. Intensity ratios of νa(CH2)/νs(CH2) for the IRRAS spectra of DPPC and the binary monolayers of DPPC and DOMA at XDOMA) 0.1 and 0.3 before and after ferritin adsorption at 30 mN/m against different angles of incidence: (a) p-polarization, (b) s-polarization.

of the PC headgroups are relatively large, the acyl chains are tilted at a tilt angle of about 30°, estimated from IRRAS21,27,46,47 and other in situ techniques,48,49 to compensate for the head-tail mismatch to form a stable monolayer at the air-water interface. The PC headgroups almost adopt a flat orientation to some extent through the electrostatic attraction between the N+(CH3)3 group of one PC headgroup and the PO2- group of its neighboring headgroup to diminish the probable electrostatic repulsion between them. In the binary monolayers, the PC headgroups are most likely to be far from the water surface to reduce the occupied surface areas due to the favorable electrostatic interaction between the PO2- groups and DOMA headgroups, taking into account the sizes of PC headgroups and the chain orientation of binary monolayers. In the binary monolayers, the -N+(CH3)2- groups of DOMA interact tightly with the PO2- groups of DPPC, leaving the N+(CH3)3 groups of DPPC to possess effective positive charges available for negative-charged protein. The protein binding mechanism is factually different from that for the binary monolayers of nonionic methyl stearate (SME) and DOMA on water free of salt,6,50 which is schematically represented in Figure 6. (46) Gericke, A.; Flach, C. R.; Mendelsohn, R. Biophys. J. 1997, 73, 492. (47) Dyck, M.; Kerth, A.; Blume, A.; Lo¨sche, M.; J. Phys. Chem. B 2006, 110, 22152. (48) Vaknin, D.; Kjaer, K.; Als-Nielsen, J.; Lo¨sche, M. Biophys. J. 1991, 59, 1325. (49) Brezesinski, G.; Dietrich, A.; Struth, B.; Bo¨hnn, C.; Bouwman, W. G.; Kjaer, K.; Mo¨hwald, H. Chem. Phys. Lipids 1995, 76, 145. (50) Dhruv, H.; Pepalla, R.; Taveras, M.; Britt, D. W. Biotechol. Prog. 2006, 22, 150.

Protein-Directed Assembly of Binary Monolayers

Figure 6. Schematic illustration of orientation of hydrocarbon chains and PC headgroups in DPPC and DPPC-DOMA monolayer together with protein binding in comparison with the SME-DOMA monolayer.

Protein-Directed Assembly of Binary Monolayers. Figure 7a-c shows SPR sensorgrams of multiple ferritin adsorption and desorption processes on the surfaces of initially fluid and preimmobilized binary monolayers at different XDOMA, respectively. The amounts of bound ferritin on the two kinds of monolayer surfaces at saturation during the first binding stage are different. At XDOMA ) 0.1, the adsorbed amount in angle shift at the initially fluid monolayer at saturation is 37% greater than that at the preimmobilized monolayer during the first binding stage, and at XDOMA ) 0.2, the adsorbed amount at the initially fluid monolayer is 27% higher than its counterpart. At a given molar ratio, the two kinds of monolayers have the same molecular densities. The increase in adsorbed amount indicates that the

Langmuir, Vol. 23, No. 15, 2007 8147

initially fluid monolayer has a higher protein affinity than its counterpart. It is most likely that the high protein affinity results from the redistribution of the binary components delivered through the lateral reorganization of lipids in the monolayer at the airwater interface, which is induced by the negative-charged ferritin in the subphase. The protein-directed assembly of the binary monolayer gives rise to a new positive charge pattern in the binary monolayer matching with the negative charge distribution on the surface of ferritin. This is different from original charge pattern in the binary monolayer mainly determined by the favorable electrostatic interaction between the two components on water free of protein. The new charge pattern in the monolayer creates multiple binding sites available for protein. It is clear that protein binding is related to the distribution of the functional lipids in the monolayers as well as the mole fraction of the functional lipids. The increased amount of adsorbed ferritin results from the directed assembly of the fluid monolayer induced by the protein in the subphase and the creation of multiple interaction sites in the monolayers favorable to protein binding. The multiple interaction sites can be preserved through the horizontal LB technique for protein recognition. At XDOMA ) 0.3, the amount of adsorbed ferritin at saturation at the fluid monolayer is slightly lower than the counterpart. At the high XDOMA, the lateral reorganization of DOMA cannot cause a significant change in the multiple binding sites due to the excess amount of DOMA. At the low XDOMA, the favorable reorganization of DOMA in the DPPC monolayers gives rise to

Figure 7. SPR sensorgrams of multiple ferritin adsorption and desorption processes on the binary monolayers of DPPC and DOMA at different mole fractions of DOMA: (a) XDOMA ) 0.1, (b) XDOMA ) 0.2, (c) XDOMA ) 0.3. (d) Relative amount of bound ferritin at the first stage of adsorption saturation as a function of XDOMA: initially fluid monolayers (solid line) and preimmobilized monolayers (dashed line). Arrow a, injecting ferritin with a final concentration of 4.0 µg/mL; arrow b, rinsing with acidic water (pH 3.0); arrow c, introducing pure water (pH 5.6).

8148 Langmuir, Vol. 23, No. 15, 2007

Du et al.

Figure 8. AFM topographic images of ferritin absorbed at the initially fluid (a) and preimmobilized (b) binary monolayers of DPPC and DOMA at XDOMA ) 0.1 with the contact mode in air: surface pressure, 30 N/m; final ferritin concentration, 4.0 µg/mL; scan size, 5 µm square. (c and d) Corresponding cross-sectional profiles.

an increase in the number of multiple binding sites for protein. The relative amount of bound ferritin against XDOMA is shown in Figure 7d. The protein affinity is enhanced at the low XDOMA. No ferritin desorption is observed from the two kinds of monolayer surfaces by exchanging the ferritin-containing subphase with pure water (pH 5.6), due to the strong electrostatic binding; however, the adsorbed ferritin is rinsed by acidic solutions with the addition of HCl (pH 3.0, the zwitterionic PC headgroups remained in this case51-53). Prior to reintroducing ferritin into the subphase, the rinsing solution of pH 3.0 is exchanged with pure water. Upon reintroducing ferritin into the subphase, the total amounts of adsorbed proteins at the preimmobilized monolayers are comparable to those during the first binding stage, and the adsorbed amounts at the initially fluid monolayers at XDOMA ) 0.1 and 0.2 are reduced to be comparable to their counterparts during the rebinding stages. For the initially fluid monolayers at XDOMA ) 0.1 and 0.2, the rebinding rates are significantly enhanced in comparison with those during the first binding ones, which indicates that the multiple interaction sites created in the initially fluid monolayers are preserved for subsequent protein binding. That means that the binary monolayers through the protein-directed assembly have a higher affinity for protein than their counterparts. Surface Patterns of Protein on the Monolayers. Figure 8a,b presents AFM topographic images of ferritin adsorbed at the initially fluid and preimmobilized monolayers at XDOMA ) 0.1 in air, and the corresponding cross-sectional profiles are shown (51) Papahadjopoulos, D. Biochim. Biophys. Acta 1968, 163, 240. (52) Zajac, J.; Chorro, C.; Lindheimer, M.; Partyka, S. Langmuir 1997, 13, 1486. (53) Gong, K.; Feng, S.-S.; Go, M. L.; Soew, P. H. Colloids Surf. A 2002, 207, 113.

in Figure 8c,d, respectively. In general, apparent widths of adsorbed molecules are much larger than their actual sizes with probes of finite sharpness, and measured heights of adsorbed molecules are comparable to their actual sizes.54 At the surface of the initially fluid monolayer, great amounts of small clusters are observed, and the corresponding step height between the adsorbed clusters and surface is about 8 nm. At the surface of the preimmobilized monolayer, larger clusters are observed, and the step height is about 12 nm. The diameters of nearly spherical ferritin molecules in aqueous solution are 12.5 nm, and the vertical dimensions of adsorbed ferritin on a bare gold surface were in the vicinity of 10 nm.55 The roughness of silicon surfaces was reported to be 0.35 nm after cleaning with organic solvents and 0.2-0.42 nm after etching by HF solution with time.56 In fact, protein flexibility, protein spreading, tip pressure, and surface roughness have an influence on the measured height of adsorbed protein. Obviously, the observed clusters on the two surfaces are adsorbed ferritin molecules at the monolayers. The protein clusters do not grow vertical to the substrate but rather grow along the in-plane direction. The driving force for this can be the van der Waals force of the clusters and the electrostatic interaction57 between the ferritin molecules and the monolayers. However, the surface patterns of the adsorbed ferritin at the two monolayer surfaces are different. The great amounts of relatively dispersed small clusters at the initially fluid monolayers are most likely (54) Johnson, C. A.; Yuan, Y.; Lenhoff, A. M. J. Colloid Interface Sci. 2000, 223, 261. (55) Caruso, F.; Furlong, D. N.; Kingshott, P. J. Collois Interface Sci. 1997, 186, 129. (56) Grisaru, H.; Cohen, Y.; Aurbach, D.; Sukenik, C. N. Langmuir 2001, 17, 1608. (57) Muguruma, H.; Kase, Y.; Murata, N.; Matumura, K. J. Phys. Chem. B 2006, 109, 26033.

Protein-Directed Assembly of Binary Monolayers

Langmuir, Vol. 23, No. 15, 2007 8149

preimmobilized monolayer (XDOMA ) 0.3) in water and the corresponding cross-sectional profile. The adsorbed ferritin clusters are uniformly distributed on the monolayers and the vertical height of adsorbed ferritin is about 12 nm. In comparison with the preimmobilized monolayer at XDOMA ) 0.1, the surface patterns of the adsorbed protein at XDOMA ) 0.3 are different. The relatively large amount of DOMA might result in the formation of multiple interactions between the adsorbed ferritin and monolayer surface, regardless of reorganization of lipid components in the monolayer. The ferritin-bound transferred monolayer was processed under water during the period of preparation and imaging, which prevents protein denature and rearrangement.

Conclusions

Figure 9. AFM topographic image of ferritin absorbed at a preimmobilized binary monolayer of DPPC and DOMA at XDOMA ) 0.3 with the contact mode in water and the corresponding crosssectional profile: surface pressure, 30 N/m; final ferritin concentration, 4.0 µg/mL; scan size, 5 µm square.

related to protein binding through the multiple interactions between the protein molecules and the reorganized monolayers. At the preimmobilized monolayer, the formation of large clusters is probably related to the rearrangement of loosely adsorbed ferritin molecules mostly through the monovalent interaction at first on the surfaces, most likely driven by the van der Waals force and hydrophobic interaction of the protein molecules in air. From the above SPR results, at XDOMA ) 0.3, monolayer reorganization has insignificant influence on protein adsorption due to the relatively large amount of DOMA. Figure 9 presents the topographic image of ferritin adsorbed at the initially

The zwitterionic DPPC monolayer at the surface pressure 30 mN/m can prevent ferritin from binding. A significant protein binding to binary monolayers of DPPC and DOMA occurs most likely through the electrostatic interaction. The adsorbed protein undergoes some change in secondary structure; however, the conformation change is inhibited due to the existence of the monolayers to some extent. At XDOMA ) 0.1, the hydrocarbon chains decrease in tilt angle from 15° ( 1° before protein binding to 10° ( 1° after protein binding, while at XDOMA ) 0.3, the chains remain vertical to the water surface before and after protein binding. In addition, protein binding to the monolayers induces somewhat chain twist with the C-C-C planes slightly parallel to the barriers at XDOMA ) 0.1 and perpendicular to the barriers at XDOMA ) 0.3. A new mechanism for protein binding to the binary monolayers is proposed. The reorganization of lipid components in the monolayers at the air-water interface, induced by the protein in the subphase, creates multiple binding sites for the protein. The generated binary surfaces through the proteindirected assembly have a high protein affinity. The created multiple binding sites in the monolayers can be preserved for use in protein recognition. Acknowledgment. The work was supported by the National Natural Science Foundation of China (Grant Nos. 20673051, 20303008, 20635020, and 20633010). Supporting Information Available: IRRAS spectra of the binary monolayers of DPPC and DOMA at the surface pressure of 30 mN/m after ferritin adsorption saturation at different angles of incidence. This material is available free of charge via the Internet at http:// pub.acs.org. LA700955F