Langmuir 2007, 23, 4629-4633
4629
Embedding of Individual Ferritin Molecules in Large, Self-Supporting Silica Nanofilms Shigenori Fujikawa,*,†,‡ Emi Muto,† and Toyoki Kunitake† Topochemical Design Laboratory, and InnoVatiVe Nanopatterning Laboratory, RIKEN, Wako, Saitama 351-0198, Japan ReceiVed December 6, 2006. In Final Form: January 20, 2007 We report herein the fabrication of ferritin-embedded self-supporting silica nanofilms via a simple spin-coating process. Ferritin was employed as a template molecule, and solutions of ferritin and silica were spread on a polymercoated silicon substrate, in this order. After dissolving the polymer underlayer by simply immersing ethanol, a centimetersized self-supporting nanofilm of ferritin/silica composite with a thickness of 15 nm was successfully transferred onto an alumina membrane without the film breaking. Ozone and hydrochloric acid solution treatment removed the template ferritin molecules from the composite film to produce corresponding transmembrane nanoholes. The reported method is very simple, and the fabrication of a protein-embedded self-supporting nanofilm enables the design of biomembranemimetic devices.
Introduction Biomembranes have unique properties as structural materials, such as self-supporting properties, structural flexibility, and molecularly thin organization. They hold proteins in the interior and on the surface of the bilayer structure to express characteristic membrane functions such as selective ion transport, signal transduction, photosynthesis, and energy conversion. Unfortunately, the fragility of biomembranes at the macroscopic scale makes their mechanical manipulation rather difficult, in spite of the attractive practical applications of their sophisticated molecular functions. Therefore, there is a strong need to develop selfsustaining nanofilms with sufficient mechanical strength as nanometer-thick matrices. Although many efforts have been devoted to fabricate free-standing nanofilms,1-7 their lateral sizes have remained a few millimeters, restricting their practical applications. Recently, we developed a facile spin-coating process for the preparation of self-supporting ultrathin films of metal oxides and metal oxide/polymer composites.8-11 Using this technique, macroscopically uniform, self-supporting metaloxide-gel films with thicknesses of several tens of nanometers and lateral dimensions of several centimeters were obtained. This self-sustaining nanofilm will be a superior candidate as a structural matrix to embed proteins in order to mimic a biological membrane. Here, we report the facile preparation of ferritin-embedded, self-supporting silica nanofilms a centimeter in size. Ferritin is * Corresponding author. E-mail:
[email protected]. † Topochemical Design Laboratory. ‡ Innovative Nanopatterning Laboratory. (1) Jiang, C.; Markutsya, S.; Pikus, Y.; Tsukruk, V. Nat. Mater. 2004, 3, 721-727. (2) Mallwitz, F.; Laschewsky, A. AdV. Mater. 2005, 17, 1296-1299. (3) Huck, W. T.; Stroock, A. D.; Whitesides, G. M. Angew. Chem., Int. Ed. 2000, 39, 1058-1061. (4) Mamedov, A.; Kotov, N. Langmuir 2000, 16, 5530-5533. (5) Xu, H.; Goedel, W. Langmuir 2002, 18, 2363-2367. (6) Jiang, C.; Markutsya, C.; Tsukruk, V. AdV. Mater. 2004, 16, 157-161. (7) Markutsya, S.; Jiang, C.; Pikus, Y.; Tsukruk, V. AdV. Funct. Mater. 2005, 15, 771-780. (8) Hashizume, M.; Kunitake, T. Langmuir 2003, 19, 10172-10178. (9) Hashizume, M.; Kunitake, T. RIKEN ReV. 2001, 38, 36-39. (10) Hashizume, M.; Kunitake, T. Soft Matter 2006, 2, 135-140. (11) Vendamme, R.; Onoue, S.; Nakao, A.; Kunitake, T. Nat. Mater. 2006, 5, 494-501.
one of the large globular proteins, with the diameter of 12 nm,12 and contains an iron oxide core (diameter, 6 nm) in its body. Both of these structural characteristics are convenient for the observation of the individual molecule by electron microscopy. Experimental Section A schematic representation of the experimental procedure is shown in Figure 1. An ethanol-soluble polymer (Tokyo Ohka Kogyo, TDURP015) was used as the underlayer, and a polymer-coated Si wafer (polymer layer, ca. 500 nm) was provided from Tokyo Ohka Kogyo (Figure 1-1). Aqueous poly(vinyl alcohol) (PVA) (Polyscience, 98 mol % hydrolyzed, Mw ∼ 78 000, 10 mg/mL) was spin-coated on the coated wafer at 3000 rpm for 2 min (Figure 1-2). In order to make the surface hydrophilic, ozone treatment was carried out by UV irradiation (UV power, 4.5 W × 3 at 254 and 180 nm) for 15 min under an oxygen pressure of 0.02 MPa on an NL-UV 253 ozone cleaner (Nippon laser & electronics lab) (Figure 1-3). Ferritin (Fluka, from horse spleen, Fe-saturated ferritin, sterile filtered solution in 150 mM NaCl) was purified by the following procedure. First, 1 mL of the commercial solution was centrifuged at 4 °C and 5000 rpm for 30 min. The precipitate was added to 1 mL of ion-exchanged water, and the mixture was again subjected to ultracentrifugation. This centrifuge process was repeated 3 times to finally obtain 2 mL of a ferritin solution, which was then spincoated on the PVA-coated substrate at 3000 rpm for 2 min (Figure 1-4). A silica precursor solution was separately prepared in the conventional two-step procedure. First, 5.58 mL of tetraethoxysilane (TEOS, Chisso Corp.), 15.4 mL of ion-exchanged water, and 2.5 mL of hydrochloric acid (4 mM) were mixed and stirred for 1 h at 70 °C. Then, 1.5 mL of hydrochloric acid (60 mM) was added, and the resulting sol was stirred for additional 1 h at 80 °C. Finally, this solution was diluted by adding ion-exchanged water to adjust the final silica concentration to 100 mM. This silica precursor was spincoated on the Si substrate at 3000 rpm for 2 min after ferritin coating (Figure 1-5). The substrate was immersed in ethanol to dissolve the polymer underlayer in order to detach the silica layer from the Si wafer (Figure 1-6). The detached self-supporting film was transferred onto a porous alumina membrane (Whatman, ANODISC, pore size 0.1 µm, diameter 25 mm, thickness 60 µm) and onto a silicon monoxidecoated Au transmission electron microscopy (TEM) grid (Spi (12) Yamashita, I. Thin Solid Films 2001, 393, 12-18.
10.1021/la0635247 CCC: $37.00 © 2007 American Chemical Society Published on Web 03/01/2007
4630 Langmuir, Vol. 23, No. 8, 2007
Fujikawa et al.
Figure 1. Preparative procedure of a ferritin/silica composite film and its chemical treatments.
Figure 2. Observation of SiO2 film morphology and the correlation of film thickness and TEOS concentration. (1) Macroscopic appearance of the SiO2 film after transfer onto alumina membrane. (2) Cross-sectional SEM image of the SiO2 film on the alumina membrane ([TEOS] ) 100 mM). (3) Correlation of film thickness and TEOS concentration. Supplies, silicon monoxide-coated 200Mesh Au) for scanning electron microscopy (SEM) and TEM observation, respectively (Figure 1-7). The transferred silica film was subjected to the UV-ozone treatment for 15 min to decompose the protein shell of the ferritin (Figure 1-8) and was then immersed in 1 N hydrochloric acid for 3 min to remove the iron core of the ferritin (Figure 1-9), followed by washing with a few droplets of ion-exchanged water. The surface morphology and the internal structure of the silica film was examined by TEM (JEOL, JEOL JEM-2100 F, acceleration voltage 200 kV without any staining), and SEM (Hitachi, S-5200 operated at an acceleration voltage of 10 kV). The sample was coated with platinum by an ion-sputtering coater (Hitachi E-1030, 20 mA, 30 s) for SEM observation. Image analysis of the TEM results was done by iTEM software (Olympus Soft Imaging Solutions), and the pore diameter in the TEM image was estimated as Feret’s diameter, which is the measured distance between parallel lines that are tangent to an object’s profile and perpendicular to the ocular scale.
Results and Discussion 1. Preparation of a Matrix Nanofilm. We have previously described the fabrication of free-standing nanofilms of metal oxides.8-10 Most of the commercially available metal alkoxides are effective as film precursors. It was essential to choose appropriate organic underlayers in order to detach the metal
oxide nanofilm without fragmentation into pieces. This fabrication procedure has to be adapted to protein incorporation in the current study. Since PVA is one of the most suitable polymers as an underlayer, we conducted the formation of silica nanofilm on this polymer. Single-component films of either PVA or SiO2 were first formed by spin-coating, but these spin-coated films were detached only as small fragments when the silicon wafer was immersed into ethanol. The combination of SiO2 and PVA was indispensable for obtaining mechanically stable nanofilms with large lateral sizes. A PVA layer may assist lateral cross-linking of SiO2 networks so that sufficient robustness is gained at the macroscopic scale. Figure 2-1 shows a macroscopic image of the SiO2 film as prepared from 100 mM of TEOS solution. The transferred film is uniform and defect-free. Figure 2-2 is a cross-section SEM image of this film. The thickness is uniform with 15 nm ( 5 nm. It is necessary to optimize the film thickness for ferritin incorporation. Thus, we used four different concentrations (100, 150, 200, and 300 mM) of the silica precursor solution and determined the thickness of the as-prepared silica film (without ferritin incorporation) by SEM observation. At all concentrations, SiO2 films with a lateral size of a few centimeters were formed
Embedding of Ferritin Molecules in Si Nanofilms
Langmuir, Vol. 23, No. 8, 2007 4631
Figure 3. SEM images of ferritin and the corresponding pore: (1) ferritin coated on a PVA-coated Si wafer, (2) after silica coating on 1, (3) after UV-ozone treatment of 2, (4) after treatment with 1 N hydrochloric acid.
without defects, and, when transferred onto alumina membranes, the film maintained the whole membrane shape without any fragmentation. The film thicknesses prepared from 100, 150, 200, and 300 mM precursors were 15, 20, 25, and 35 nm (error range ( 5 nm in all cases), respectively, and almost linearly increased in proportion to the precursor concentration. Its slope was approximately 10 nm/100 mM. We also investigated the film thickness prepared from lower TEOS concentrations (for example, 50 mM). However, a uniform film was not obtained, and the transfer was difficult. The 100 mM TEOS gave a film thickness comparable to the diameter of ferritin, and this concentration was employed in the following embedding experiment. 2. Ferritin Embedding. Ferritin embedding was examined with a silica nanofilm of ca. 15 nm thickness. This was the smallest thickness we could manipulate as large, free-standing film. When the aqueous solution of ferritin molecules was spincoated on the PVA-coated substrate, nanoparticles (conceivably ferritin) were abundantly observed on the substrate by SEM (Figure 3-1). The diameter of the individual particle was 15-20 nm, and was close to that of a ferritin molecule (12 nm) if we take into account the effect of Pt sputtering for the SEM observation. The additional spin-coating of silica precursors produced a smooth surface again, and the particles mostly disappeared, although a small number of them were observed (Figure 3-2). Ferritin proteins may have been buried in the silica film, or the second spin-coating of the silica solution may have washed away the ferritin particles. In order to discriminate these two possibilities, ozone treatment was applied to remove the organic moiety from the ferritin-embedded silica film. Cracks and holes developed after ozone treatment (Figure 3-3). Such cracks and holes were not observed in the absence of ferritin in the silica film. A small number of the holes observed by SEM (pointed by the arrow in Figure 3-3) had diameters of less than 20 nm, in fair agreement with the size of a ferritin molecule. The hole size may be affected by the presence of the sputtered platinum for SEM observation (usually 2 nm thickness for coating) and by unremoved iron particles of ferritin. The ozone-treated silica film was immersed in the dilute hydrochloric acid solution to remove iron oxide particles. The hole structure became more clearly observable by SEM (Figure 3-4, indicated by the arrows), indicating the presence of buried ferritin.
Direct observation of ferritin molecules buried in the silica film is difficult by SEM since we cannot examine the film interior. TEM observation is better suited for this purpose. Thus, the ferritin/SiO2 composite film was detached from the substrate by dissolving the polymer underlayer and was transferred onto the TEM grid without any fragmentation. In the TEM observation, black dots with a diameter of several nanometers were abundantly observed without overlayering (Figure 4-1). This specimen was not stained, and the black dots were not observed in a silica film without ferritin. Therefore, these dots must be attributed to the iron core of ferritin. The ferritin molecules are fully dispersed in the film, in agreement with SEM observation. The black dots were unchanged even after UV-ozone treatment (Figure 4-2), but the dots were slightly overlapped. The pore structures are more clearly seen in the film after acid treatment (Figure 4-3). The number of black dots (i.e., iron core) is obviously lessened, but they remain within larger pores. Undoubtedly, the iron core that existed in the pore was mostly dissolved away. The pore size was estimated from the film after acid treatment; clearly observable pores, as indicated by the dotted circles in Figure 5a, were used for the image analysis. The pore size was estimated as Feret’s diameter. The measured pore size gave a distribution as shown in Figure 5b, and the average size of 11. 9 nm is nearly identical to the diameter of the ferritin molecule.12 The precise comparison of the protein structure and the pore shape provides more direct evidence of the imprint effect for the pore formation. Unfortunately, the resolution of the electron microscopic technique we employed was not sufficient to perform such a precise shape comparison. Therefore, we conclude that the individual ferritin molecule acted as a template to form the corresponding nanopore. The metal oxide network must be flexible enough to adjust their network structure to the template morphology without losing mechanical robustness to maintain macroscopic film shape. The thickness of this silica film matrix is estimated to be 15 ( 5 nm, and the diameter of a ferritin protein is known to be 12 nm.12 Electron microscopy indicates that the protein molecule is dispersed in the matrix individually. From these results, we propose the structure model in Figure 6 for the composite nanofilm. Protein-embedded membrane systems have been reported from many research groups. Polymer-supported lipid bilayer membranes have been studied as models of the cell surface that mimic many of the physicochemical properties of a biological
4632 Langmuir, Vol. 23, No. 8, 2007
Fujikawa et al.
Figure 4. TEM images of ferritin-embedded SiO2 film: (1) as-prepared ferritin-embedded SiO2 film, (2) after UV-ozone treatment, (3) after treatment with 1 N hydrochloric acid, (4) a magnified image of the dot line box in image 3.
Figure 6. A structural model of a ferritin-embedded SiO2 film.
Figure 5. The pore diameter in the SiO2 film. (a) The pore used for analysis is surrounded by a dotted circle. (b) The distribution of the pore diameter.
membrane.13-15 We also reported the layer-by-layer assembly of proteins and electrolyte polymers,16 and Ratner et al. described the preparation of a surface-imprinted membrane for protein recognition.17 In the latter case, protein molecules were employed as a template to fabricate nanopits on the substrate surface, and the as-prepared nanopit could recognize the template protein. (13) Tanaka, M.; Sackmann, E. Nature 2005, 437, 656-663. (14) Parikh, A. N.; Groves, J. T. MRS Bull. 2006, 31, 507-509. (15) Tanaka, M. MRS Bull. 2006, 31, 513-520. (16) Lvov, Y.; Ariga, K.; Ichinose, I.; Kunitake, T. J. Am. Chem. Soc. 1995, 117, 22, 6117-6123. (17) Shi, H.; Tsai, W.-B.; Garrison, M. D.; Ferrari, S.; Ratner, B. D. Nature 1999, 398, 593-597.
Unfortunately, the transmembrane function cannot be realized by these solid-supported membrane systems. Recently more elaborate systems were developed by using polymer cushions or by attaching lipopolymer tether on the lower side of supported flat bilayer membranes.18-23 In our system, both ends of the protein molecule are essentially exposed to the upper and lower surfaces of the nanofilm, which is separable as a self-supporting film. Thus, the embedded protein molecule is accessible from both sides of the film, and the transmembrane functions may be realized. The spatial orientation of the protein is not controlled in the current fabrication procedure, and the electron microscopic technique we employed is not effective for determining the orientation. In this case, it is noteworthy that the present, robust film can be manipulated macroscopically. This characteristic is advantageous for the design and detection of anisotropic transmembrane properties. The dynamic property of lipid bilayer membranes cannot be reproduced by our silica film. Recently, we reported that robust, self-supporting nanomembranes were (18) Lang, H.; Duschl, C; Vogel, H. Langmuir 1994, 10, 197. (19) Cornell, B. A.; Braach-Maksvytis, V.; King, L. G.; Osman, P. D. J.; Raguse, B.; Wieczorek, L.; Pace, R. J. Nature 1997, 387, 580. (20) Schiller, S. M.; Naumann, R.; Lovejoy, K.; Kunz, H.; Knoll, W. Angew. Chem., Int. Ed. 2003, 42, 208. (21) Wagner, M. L.; Tamm, L. K. Biophys. J. 2000, 79, 1400. (22) Bunjes, N.; Schmidt, E. K.; Jonczyk, A.; Rippmann, F.; Beyer, D.; Ringsdorf, H.; Gra¨ber, P.; Knoll, W.; Naumann, R. Langmuir 1997, 13, 6188. (23) Purrucker, O.; Fro¨tig, A.; Jordan, R.; Tanaka, M. ChemPhysChem 2004, 5, 327.
Embedding of Ferritin Molecules in Si Nanofilms
fabricated from organic/inorganic hybrid materials. Appropriate design of such hybrid nanofilms may satisfy the robustness and fluidity required for lipid bilayer mimetics.
Conclusion Size matching of a film thickness and component molecules enables the exploitation of molecule-based events in the form of macroscopic material features. Biomembranes are a typical example of sophisticated functions that are derived from the size matching of a nanolayer (lipid bilayer) and a nano-object (protein). Similarly, nanometer-thick and self-supporting
Langmuir, Vol. 23, No. 8, 2007 4633
films will lead to novel transmembrane phenomena in artificial systems. Although we used a single, globular protein as a template to test the feasibility of our approach, more complex units of functional proteins may be readily incorporated into a nanofilm in a similar manner. Macroscopically self-supporting protein/ nanofilm composites have great potential for a variety of applications. In conclusion, it is clear that the current approach offers great potential for the fabrication of biomimetic nanomembranes. LA0635247
