- Email: [email protected]
Characterization of silica-coated tobacco mosaic virus
Journal of Colloid and Interface Science 298 (2006) 706–712 www.elsevier.com/locate/jcis
Characterization of silica-coated tobacco mosaic virus Elizabeth Royston a , Sang-Yup Lee a , James N. Culver b , Michael T. Harris a,∗ a School of Chemical Engineering, Purdue University, West Lafayette, IN 47907, USA b Center for Biosystems Research, University of Maryland Biotechnology Institute, College Park, MD 20742, USA
Received 30 September 2005; accepted 30 December 2005 Available online 19 January 2006
Abstract The deposition of silica on the surface of tobacco mosaic virus (TMV) is achieved at a higher pH (>7) as a means to enhance its usefulness as a template for the synthesis of nanostructures. Electron energy loss spectroscopy definitively shows the presence of a silica shell on the surface of the TMV while small angle X-ray scattering differentiates successfully between silica-coated TMV and silica particles in the presence of uncoated TMV. Importantly, coating reactions done in a 50% w/v methanol/water solution produce smaller silica nanostructures during the condensation of the hydrolysis intermediates, possibly aiding in obtaining uniform coating. Furthermore, TMV-templated silica coatings are found to enhance the stability of the virus particle in methanol at conditions that would ordinarily disrupt the assembled particle. Combined these findings demonstrate that TMV can function as an efficient template for the controlled deposition of silica at neutral pH. © 2006 Published by Elsevier Inc. Keywords: Tobacco mosaic virus; Silica, biotemplate; EELS; SAXS
1. Introduction Preparation of nanoscaled materials with complex structures, such as nanowires and nanotubes, is of great importance in regard to miniaturization. Currently there are several methods of creating nanowires [1–5] including the use of biological structures, such as virions [6–9], as templates for the synthesis of specific nanostructures. Tobacco mosaic virus (TMV) produces a tube-like virus particle, 300 nm long by 18 nm wide, which is stable over a wide range of pH and temperatures, thus making it an ideal nanostructure template. The coating of various metals and metal sulfide nanoparticles on the surface of wild-type and genetically modified TMV have been previously reported [10–15]. Additionally, silicon dioxide gels (from tetraethylorthosilicate (TEOS) with a strong acid in ethanol) formed on wild-type TMV have been presented by researchers, along with the creation of mesoporous silica structures formed using a wild-type TMV template [10,16]. The tubular structure of the TMV particle consists of ∼2130 identical protein subunits of molecular weight 17.5 kDa * Corresponding author. Fax: +1 765 494 0805.
E-mail address: [email protected] (M.T. Harris). 0021-9797/$ – see front matter © 2006 Published by Elsevier Inc. doi:10.1016/j.jcis.2005.12.068
wrapped in a helix around a single strand of RNA [17]. These particles are stable in solutions of approximately pH 2–10 and at temperatures up to 60 ◦ C [18]. The surface of the assembled virus particle contains numerous glutamic acid, aspartic acid, lysine and arginine residues, giving an overall negative surface charge above the isoelectric point, pH 3.0–3.5, or a positive charge when below. The length of the TMV particle is controlled by the length of the encapsulated viral RNA. However, a genetically modified TMV mutant, E50Q, contains a glutamic acid to glutamine mutation at position 50 within the coat protein open reading frame. This mutation neutralizes a negative–negative charge repulsion involved in the controlled assembly of the virion. As a result, the E50Q mutant produces virus particles that assemble beyond the length of the encapsulated RNA, averaging approximately 1000 nm in length yet still maintaining the properties of the wild-type virus [19]. Three pH domains for the polymerization of silica determine the morphology of the silicate oxide products: pH < 2, pH 2–7, and pH > 7 [20]. Previous examples of silica coatings created by Shenton et al. were conducted in strong acidic conditions (pH 2.5) where the acid catalyzed hydrolysis of TEOS in an ethanol solvent (between the isoelectric points of TMV (pI 3.5) [21] and silica (pI 2) [22]) produced a charge–charge
E. Royston et al. / Journal of Colloid and Interface Science 298 (2006) 706–712
interaction between the negatively charged silica and the positively charged TMV surface [10]. However, a search for milder reaction conditions in addition to a more detailed characterization of the silica-coated TMV is needed for the future application of TMV-based silicate oxide nanotubes. Milder reaction conditions at the slightly basic conditions are not only safer but provide greater control over gelation by slowing the overall gelation process in comparison to a similar reaction system under acidic conditions. At low ionic concentration (<2 mM monovalent ions) and low TEOS concentration, electrostatic repulsion plays a prominent role in slowing the gelation process above pH 8 for aqueous systems [20]. Furthermore, the rate of hydrolysis reduces while the rate of silica dissolution increases in higher pH conditions [22] resulting in an overall decrease in the rate of gelation. The research presented in this paper therefore achieves silica layer formation in aqueous methanolic solutions in a pH > 7. The hydrolysis and condensation reactions of TEOS in aqueous methanolic solutions produce denser nucleating particles about half the diameter of those formed under similar reaction conditions in ethanol [23]. These particles may help produce smoother coatings with increased density as a result of the increased mobility and smaller negatively-charged surface area of the smaller silica particles. Due to stability concerns with the virus in alcohol–water solutions, the concentration of water must be at least 40 wt%. Alcohol denatures proteins by disrupting the intramolecular hydrogen bonding, where new hydrogen bonds are formed instead between the alcohol molecule and the protein side chains [24]. Thus for purposes of this paper, concentrations of alcohol– water mixtures are maintained at 50–50 wt%. Characterization of the silica-TMV nanotubes is performed using electron energy loss spectroscopy (EELS), small angle X-ray scattering (SAXS), and dynamic light scattering (DLS). EELS provides high spatial resolution elemental analysis of the silica coating, while SAXS and DLS provide in situ information on the structures of the system. The detailed structural information resulting from this research will be useful in utilizing silica-TMV nanotubes in other applications, such as multi-layered coatings. 2. Experimental 2.1. Materials Tetraethylorthosilicate (Aldrich 98%) and ammonium hydroxide (NH4 OH, Aldrich 5N) were used as received. HPLC grade methanol (MeOH, 99%) and deionized water were used in all preparations. Both the E50Q mutant and wild-type TMV were prepared as previously described [19]. 2.2. Coating TMV A 50–50 wt% solution of alcohol to water was used as the solvent. The base catalyst was added to a total concentration of 0.001 M NH3 , followed by TMV to a total concentration of 0.2 mg/mL, and finally by TEOS to a total concentration of 0.009 M. The NH3 concentrations were kept low due to stability concerns with the TMV. The solution was shaken vigorously
707
between additions of each reactant. The reaction mixture was allowed to stand for 24 h to ensure successful growth of a silica shell around the virus. Reference buffer solutions for pH measurements in a 50–50 wt% methanol–water mixture were prepared as described by Paabo et al. [25]. The pH of the sample was measured at the start and end of the reaction and read at pH∗ = 8.8 for both reaction times. This result was expected as the change in pH due to the silica reaction should be small as the concentration of TEOS in solution is very low. To remove the excess silica from solution once the shell has been created, the sample was centrifuged at 12,500 rpm for 15 min several times and re-suspended in methanol. TEM images of the solution were taken after 24 h. 2.3. Characterization Transmission electron microscopy (TEM) was conducted using a Philips CM10 microscope. The particle size was verified from measurements taken from several negatives of each sample. All TEM samples were prepared without staining by placing a drop of sample on a carbon coated copper grid followed by wicking away the excess liquid. Electron energy loss spectroscopy (EELS) was carried out using a Vacuum Generator HB501 STEM. Cold-stage EELS studies were conducted using a JEOL 2010F TEM. The beam energies were 80 and 200 keV (100 and 200 keV max), respectively. Silica maps were taken on the 2010F using energy centers of 82, 92, and 113 eV with 10 eV spreads for 16 s, with the dried sample held at −174 ◦ C during the spectrum collection. Small angle X-ray scattering (SAXS) measurements were taken using a Molecular Metrology X-ray scattering instrument with a microfocus X-ray source, an Osmic MaxFlux confocal X-ray optic, a three-pinhole camera, and a gas-filled 2D multiwire detector at a 1424 mm camera length. Calibration in q-space was achieved with an isotropic silver behenate powder standard. The hydrodynamic radius of the particles was measured at room temperature using a dynamic light scattering (DLS) system (Photocor Instruments) that was equipped with a 20 mW He–Ne laser (λ = 632 nm). All measurements were taken at a 90◦ scattering angle. 3. Results and discussion TEM images of unstained silica-coated wild-type TMV and E50Q mutant TMV particles re-suspended in 100% methanol are shown in Figs. 1a and 1b, respectively. The inserts in Figs. 1a and 1b show uncoated and unstained wild-type TMV and E50Q in water. In Fig. 1a, the arrows indicate the length of one virion, since the image shows two virions lined end to end. The lengths of the nanotubes are consistent with those of wild-type TMV (300 ± 29 nm) and E50Q mutant TMV (940 ± 470 nm) templates [19]. Due to the low electron density of silica, the thin silica shell on the TMV is difficult to observe clearly. TEM observation at higher magnification re-
708
E. Royston et al. / Journal of Colloid and Interface Science 298 (2006) 706–712
(a)
(a)
(b)
Fig. 2. (a) DLS results for silica-coated wild-type TMV (1) in reaction conditions over time. (b) DLS results for silica-coated E50Q (P) in reaction conditions over time. The bars show one standard deviation, the points are the means. The shaded areas are the calculated Rh ranges.
(b) Fig. 1. (a) TEM image, unstained, of silica-coated wild-type TMV suspended in 100% MeOH. (b) TEM image, unstained, of silica-coated E50Q suspended in 100% MeOH. The inserts are unstained and uncoated wild-type TMV and E50Q respectively. In (a) the arrows indicate the length of one virion, since the image shows two virions lined end to end. Scale is the same between each insert and image.
sults in blurring of the image and failure to show a discrete silica shell. DLS studies help to show that the dispersion and stability of the virus particles remain unaffected by the coating process. The hydrodynamic radius of the silica-coated wild-type TMV before centrifugation and re-suspension in methanol is found to be around 55 nm, whereas the E50Q mutant has a hydrodynamic radius of around 60 nm, consistent with the presence of longer virion particles. Fig. 2 shows the Rh of the silicacoated TMV particles over time. From the theoretical diffusion coefficient for a cylinder [26,27] and the given values for the distribution of sizes for the uncoated virus, the approximate expected value of the Rh of wild-type TMV and E50Q falls between 50–60 nm and 70–160 nm, respectively. The experimental DLS values for Rh are in agreement with the theoretical values, as shown by the shaded areas in Fig. 2. Although there is significant overlap of the error bars, the mean Rh values for wtTMV and E50Q are statistically different at approximately the 90% confidence level. The smaller size range for the E50Q particles is likely a result of a combination of factors, such as fragmentation of the virus during the coating process due to the alcohol and ammonia in the solution and inherent length differences associated with different E50Q preparations. The increase in length of the silica-coated E50Q as compared to wtTMV agrees with TEM observations, as seen in Figs. 1a and 1b.
TMV relies on water bridging to maintain the conformational stability of its proteins [28]. Mixtures with concentrations higher than 60 wt% alcohol compromise the stability of the TMV structure. In these situations, the bridging water molecules are reoriented and displaced by alcohol molecules and the TMV disassembles. The results presented in this paper demonstrate stable silica-coated TMV particles in solutions of 100% methanol, indicating the presence of another “force” stabilizing the rod-like structure of the TMV. It is likely that the silica shell on the TMV surface prevents the reorientation of the water molecules, thus keeping the alcohol groups from binding with the protein. Re-suspension of the silica-coated TMV in water results in the breakup of the TMV structure over extended periods of time. This is most likely due to the increased solubility of silica in water versus the solubility of silica in methanol. Silica dissolves off the surface of the TMV and causes instability of the protein structure as it leaves. The exact nature of this instability has not been determined. TEM was used to monitor the effect of time on re-suspensions of silica-coated TMV in methanol, water, and silica-saturated water. After 1 h, silicacoated TMV in water shows definite signs of breakup, while the silica-coated TMV in methanol and silica-saturated water both contain stable virions. After 3 h, there is no sign of TMV in water, whereas unchanged stable virions are still seen in methanol and silica-saturated water solutions. The TEM results for samples suspended in methanol and silica-saturated water at 3 h and the result for a sample suspended in water at 1 h are shown in Fig. 3. No image is included for the silica-coated TMV re-suspended in pure water at 3 h as there are no TMV particles present on the grid. This result suggests that the association of silica with the TMV coat protein changes the nature of the molecular interaction between neighboring coat proteins in some fashion. Electron energy loss spectroscopy (EELS) and cold-stage EELS are used to confirm the presence and location of silica
E. Royston et al. / Journal of Colloid and Interface Science 298 (2006) 706–712
709
(a)
(a) (b)
(c) Fig. 3. TEM images of silica-coated TMV re-suspended in (a) MeOH after 3 h (b) silica-saturated H2 O after 3 h (c) H2 O after 1 h. An image of silica-coated TMV re-suspended in H2 O is not included as there are no particles present on the TEM grid. The scale bar is 300 nm in all images.
and TMV in the sample. EELS is chosen for its high spatial resolution (as low as 0.3–0.5 nm in some cases), unlike X-ray emission spectroscopy which suffers from beam broadening due to backscattering, secondary fluorescence, and the generation of fast secondary electrons in the specimen [29]. Scanning transmission electron microscopy (STEM) is used in conjunction with EELS, while TEM is used with the cold-stage EELS. Previous EELS studies conducted on TMV by Leapman et al. show the presence of carbon and phosphorous (from the RNA) indicate the existence of TMV [30]. Figs. 4a and 4b (EELS results) show the presence of phosphorous and additional carbon on the rods relative to the background, indicating the rod structures are TMV. The log-differential of the EELS signal intensity is plotted in order to show the peaks that are otherwise difficult to discern (Fig. 4b). To verify the presence of the silica on the TMV, a silica map of the silica-coated TMV sample is made. Cold-stage EELS is used in order to minimize the effects of beam damage present
(b) Fig. 4. (a) Darkfield TEM image of silica-coated wild-type TMV. Silica appears in the background as bright clusters. (b) Log-differential plot of the EELS data displaying the peaks with clarity. The data corresponds to the areas indicated by the arrows in (a).
in a regular EELS sample holder. The silica map of the silicacoated wild-type TMV re-suspended in methanol is seen in Fig. 5b. The average diameter of the TMV is 19 ± 0.5 nm. Results in Figs. 5a and 5b show an evenly distributed silica presence over the entire surface of the TMV, which would not be the case if this was driven by capillary forces during drying. In that case, the silica would demonstrate a higher presence along the outside edges of the TMV image, creating an outline of electron dense material. SAXS studies clearly show the difference between the silicacoated TMV and uncoated TMV particles in the presence of preformed silica particles. The preformed silica particles are made using the same reaction condition as that used for the silica-coated TMV. The S/N ratio of the instrument is too low
710
E. Royston et al. / Journal of Colloid and Interface Science 298 (2006) 706–712
(a)
(a)
(b) Fig. 6. SAXS data demonstrating the differences between (1) the reaction of TEOS in the presence of TMV, (P) the reaction of TEOS without TMV, (E) TMV in water, and (!) TMV added to TEOS reacted without TMV present. (a) Intensity vs Q plot of the data. (b) ln(I ) vs Q2 plot showing linearity of data for Guinier analysis. (b) Fig. 5. (a) TEM image of silica coated wild-type TMV. (b) Silica map taken using cold-stage EELS (−174 ◦ C).
at the lower intensity levels to clearly show the expected smaller peaks associated with TMV [31]; therefore, the SAXS results, as shown in Fig. 6, are used to demonstrate the differences between the systems, in situ. Studies on the length of exposure to the beam and the degradation of the sample indicate no measurable damage during the 1 h sample collection time. Fig. 6a shows the scattering intensity versus Q (nm−1 ) data. The SAXS results show a decrease in scattering in the 0.2– 1 nm−1 Q range when the TEOS is reacted to form silica in the presence of TMV versus scattering when the silica is preformed and is subsequently mixed with TMV. The scattering curve for the sample with preformed silica particles and uncoated TMV is a combination of the scattering from the silica particles and the TMV in water. The decreased scattering intensity for the silica-coated TMV sample may be due to a continuous coating of small silica structures on the TMV surface. USAXS experiments are planned to determine if more intense scattering occurs at lower Q for the silica-coated TMV system,
Table 1 Rg values calculated from SAXS data over 0.5 < Q2 < 2.25
Silica-coated TMV Supernatant Concentrated re-suspension Silica particles Silica + TMV
Rg
R2
1.70 1.84 0.86 1.86 1.92
0.989 0.992 0.893 0.993 0.998
which would indicate the formation of a continuous coating of silica on the TMV particles and would account for the decrease in scattering intensity, I , in the ∼0.2 nm−1 Q range. ln(I ) versus Q2 plots are shown in Fig. 6b to determine if there is a sufficient Guinier region to obtain radius of gyration, Rg , values. The scattering intensity data in the region 0.5 < Q2 < 2.25 show a linear behavior. The results of the linear regression analysis are given in Table 1 and are discussed below. Fig. 7 shows the SAXS results detailing the effectiveness of centrifugation in separating the unattached silica particles from the silica-coated TMV particles. This step is essential in minimizing residual silica particles during the preparation of
E. Royston et al. / Journal of Colloid and Interface Science 298 (2006) 706–712
(a)
(b) Fig. 7. (a) SAXS data of (1) the reaction of TEOS in the presence of TMV, (P) TMV in water, (E) the supernatant from the separation of silica-coated TMV from the silica background, and (!) re-suspension of silica-coated TMV in MeOH. (b) ln(I ) vs Q2 plot showing linearity of data for Guinier analysis for (1) the reaction of TEOS in the presence of TMV, (E) the supernatant from the separation of silica-coated TMV from the silica background, (!) re-suspension of silica-coated TMV in MeOH, and (e) the concentrated re-suspension of the silica-coated TMV in MeOH.
silica-coated TMV samples for all TEM images. It is also important to remove the residual silica for the EELS analysis. The majority of the silica and silica-coated TMV is found in the supernatant since the scattering curves (Figs. 7a and 7b) for the silica-coated TMV sample and the supernatant almost overlap. The re-suspended pellet shows a scattering curve (Fig. 7a) containing TMV without background silica. Unfortunately, the X-ray scattering curves display a very noisy pattern (Fig. 7b) when the suspension is diluted to the original sample volume and it renders Guinier analysis impossible. Therefore, the suspension is concentrated by a factor of six over the course of the multiple centrifugations. The SAXS data for the concentrated suspension is also shown in Fig. 7b and is used in the Guinier analysis for silica-coated TMV re-suspended in MeOH. The radius of gyration (Rg ) values calculated from the linear section of the ln(I ) vs Q2 curve (slope = Rg2 /3), in addition to the R 2 values for the linear fit, are reported in Table 1. From the results, it appears that the Rg values (approximately 1.8–1.9 nm) are for the silica particles in solution. The radii of gyration for these particles are approximately half the size of the first nanophases
711
that are reported by Green et al. [23] for the hydrolysis and condensation of TEOS in aqueous-methanolic solutions and at much lower water concentrations. The decrease in the size of silica nanophase that is observed in the current experiments is most probably due to the higher water concentration (approximately 24 mol/L) during the hydrolysis and condensation of TEOS. Aggregates of silica particles are confirmed in the background of the STEM image, Fig. 4a, with the smaller particles being on the order of 2 nm in diameter. As expected, the Rg values for the silica (1.86 nm) and the silica + TMV (1.92 nm) systems are similar since, relative to the scattering from the silica particles, the scattering by TMV is very low in the Q range where the Guinier analysis is done. In other words, scattering by the silica particles dominates in this Q range. The silica nanostructures in the supernatant have an Rg value of 1.84 nm which is similar to that of the silica particle system. Thus, 1.8–1.9 nm silica nanostructures are still being formed when TEOS is reacted to form silica in the presence of TMV; however, the lower intensity of the scattering curve relative to the scattering curve for the silica + TMV system in the 0.2–1.5 nm Q range indicates that fewer of these nanostructures are present. The Rg (0.9 nm) associated with the nanostructures in the concentrated re-suspension of the silica-coated TMV are possibly due to the presence of smaller nanostructures that detach from the surface of the silica-coated TMV when the methanol is added to re-suspend the sample. The mechanism for the coating of silica on TMV by the hydrolysis and condensation of TEOS is not understood. Hydrolyzed and partially condensed species of TEOS may be attracted to the surface of the TMV due to attractive particle interactions like van der Waals forces. In addition, the negatively charged silica monomers may be attracted to the positively charged arginine and lysine amino acids on the surface of the TMV. Furthermore, studies investigating the interaction mechanism between proteins and hydrophilic silica surfaces [32] indicate that interactions between hydroxyl groups on the silica surface and carboxyl/carbonyl or amine/imido groups [33] in the protein structure should contribute to the formation of the silica shell on the TMV surface. Once the initial hydrolysis is completed and the partially condensed species are attached to the surface of the TMV, additional silica precursors may participate in condensation reactions with the attached silicates to form a gel on the surface of the TMV. Additional research is required to fully elucidate the reaction. 4. Conclusion Silica-coated TMV has been successfully produced by the reaction of TEOS in methanol and water at pH values above those that have been previously reported. Stabilization of the TMV in methanol has been achieved through the silica coating. TEM studies combined with EELS analysis confirm the presence of silica on the surface of the TMV while SAXS studies show the differences between uncoated and coated TMV particles. The E50Q mutant has also been silica-coated successfully, giving rise to longer silica nanostructures. Successful charac-
712
E. Royston et al. / Journal of Colloid and Interface Science 298 (2006) 706–712
terization of the system provides the information required to enhance the usefulness of the silica-coated TMV as a template. Acknowledgments This research is supported by the U.S. Department of Energy (DEFG02-02-ER45975 and DEFG02-02-ER45976). TEM work is done in the Life Science Microscopy Facility, Purdue University. EELS analysis is carried out in the Center for Microanalysis of Materials, University of Illinois, which is partially supported by the U.S. Department of Energy under Grant DEFG02-91-ER45439. Special thanks go to Dr. Ray D. Twesten for his assistance in collecting the EELS data. SAXS studies are supported by the NSF funded (MRI program award 0321118-CTS) facility for in situ X-ray scattering from nanomaterials and catalysts, Purdue University. References [1] A. Huezko, Appl. Phys. A: Mater. Sci. Process. 70 (2000) 365. [2] X. Gao, H. Matsui, Adv. Mater. 17 (2005) 2037. [3] R. Kirsch, M. Mertig, W. Pompe, R. Wahl, G. Sadowski, K.-J. Böhm, E. Unger, Thin Solid Films 305 (1997) 248. [4] J.L. Coffer, S.R. Bigham, X. Li, R.F. Pinizzotto, Y. Rho, G. Young, R.M. Pirtle, I.L. Pirtle, Appl. Phys. Lett. 69 (1996) 3851. [5] M. Markowitz, S. Baral, S. Brandow, A. Singh, Thin Solid Films 224 (1993) 242. [6] T. Douglas, M. Young, Adv. Mater. 11 (1999) 679. [7] Y. Huang, C.Y. Chiang, S.K. Lee, Y. Gao, E.L. Hu, J. De Yoreo, A.M. Belcher, Nano Lett. 5 (2005) 1429. [8] C.E. Flynn, S.-W. Lee, B.R. Peelle, A.M. Belcher, Acta Mater. 51 (2003) 5867. [9] C. Mao, D.J. Solis, B.D. Reiss, S.T. Kottman, R.Y. Sweeney, A. Hayhurst, G. Georgiou, B. Iverson, A.M. Belcher, Science 303 (2004) 213.
[10] W. Shenton, T. Douglas, M. Young, G. Stubbs, S. Mann, Adv. Mater. 11 (1999) 253. [11] E. Dujardin, C. Peet, G. Stubbs, J.N. Culver, S. Mann, Nano Lett. 3 (2003) 413. [12] W.L. Liu, K. Alim, A.A. Balandin, Appl. Phys. Lett. 86 (2005) 253108. [13] S. Lee, E. Royston, J.N. Culver, M.T. Harris, Nanotechnology 16 (2005) S435. [14] M. Knez, M. Sumser, A.M. Bittner, C. Wege, H. Jeske, S. Kooi, M. Burghard, K. Kern, J. Electroanal. Chem. 522 (2002) 70. [15] M. Knez, A.M. Bittner, F. Boes, C. Wege, H. Jeske, E. Maiß, K. Kern, Nano Lett. 3 (2003) 1079. [16] C.E. Fowler, W. Shenton, G. Stubbs, S. Mann, Adv. Mater. 13 (2001) 1266. [17] K. Namba, R. Pattanayek, G. Stubbs, J. Mol. Biol. 208 (1989) 307. [18] G. Stubbs, Seminar Virology 1 (1990) 405. [19] B. Lu, G. Stubbs, J.N. Culver, Virology 225 (1996) 11. [20] R.K. Iler, The Chemistry of Silica, Wiley, New York, 1979. [21] G. Oster, J. Biol. Chem. 190 (1951) 55. [22] C.J. Brinker, G.W. Scherer, Sol–Gel Science, Academic Press, New York, 1990. [23] D.L. Green, J.S. Lin, Y. Lam, M.Z.-C. Hu, D.W. Schaefer, M.T. Harris, J. Colloid Interface Sci. 266 (2003) 346. [24] A.L. Lehninger, D.L. Nelson, M.M. Cox, Principles of Biochemistry, second ed., Worth Publishers Inc., New York, 1993. [25] M. Paabo, R.A. Robinson, R.G. Bates, J. Am. Chem. Soc. 87 (1965) 415. [26] M. Doi, S.F. Edwards, The Theory of Polymer Dynamics, Oxford Univ. Press, Oxford, 1986. [27] N.C. Santos, M.A.R.B. Castanho, Biophys. J. 71 (1996) 1641. [28] A.C. Bloomer, J.N. Champness, G. Bricogne, R. Staden, A. Klug, Nature 276 (1978) 362. [29] R. Egerton, Electron Energy-Loss Spectroscopy in the Electron Microscope, second ed., Plenum Press, New York, 1996. [30] R.D. Leapman, N.W. Rizzo, Ultramicroscopy 78 (1999) 251. [31] A.G. Malmon, Acta Crystallogr. 10 (1957) 639. [32] Y.I. Tarasevich, L.I. Monakhova, Colloid J. 64 (2002) 482. [33] D. Christendat, T. Abraham, Z. Xu, J. Masliyah, J. Adhes. Sci. Technol. 19 (2005) 149.
Characterization of silica-coated tobacco mosaic virus Elizabeth Royston a , Sang-Yup Lee a , James N. Culver b , Michael T. Harris a,∗ a School of Chemical Engineering, Purdue University, West Lafayette, IN 47907, USA b Center for Biosystems Research, University of Maryland Biotechnology Institute, College Park, MD 20742, USA
Received 30 September 2005; accepted 30 December 2005 Available online 19 January 2006
Abstract The deposition of silica on the surface of tobacco mosaic virus (TMV) is achieved at a higher pH (>7) as a means to enhance its usefulness as a template for the synthesis of nanostructures. Electron energy loss spectroscopy definitively shows the presence of a silica shell on the surface of the TMV while small angle X-ray scattering differentiates successfully between silica-coated TMV and silica particles in the presence of uncoated TMV. Importantly, coating reactions done in a 50% w/v methanol/water solution produce smaller silica nanostructures during the condensation of the hydrolysis intermediates, possibly aiding in obtaining uniform coating. Furthermore, TMV-templated silica coatings are found to enhance the stability of the virus particle in methanol at conditions that would ordinarily disrupt the assembled particle. Combined these findings demonstrate that TMV can function as an efficient template for the controlled deposition of silica at neutral pH. © 2006 Published by Elsevier Inc. Keywords: Tobacco mosaic virus; Silica, biotemplate; EELS; SAXS
1. Introduction Preparation of nanoscaled materials with complex structures, such as nanowires and nanotubes, is of great importance in regard to miniaturization. Currently there are several methods of creating nanowires [1–5] including the use of biological structures, such as virions [6–9], as templates for the synthesis of specific nanostructures. Tobacco mosaic virus (TMV) produces a tube-like virus particle, 300 nm long by 18 nm wide, which is stable over a wide range of pH and temperatures, thus making it an ideal nanostructure template. The coating of various metals and metal sulfide nanoparticles on the surface of wild-type and genetically modified TMV have been previously reported [10–15]. Additionally, silicon dioxide gels (from tetraethylorthosilicate (TEOS) with a strong acid in ethanol) formed on wild-type TMV have been presented by researchers, along with the creation of mesoporous silica structures formed using a wild-type TMV template [10,16]. The tubular structure of the TMV particle consists of ∼2130 identical protein subunits of molecular weight 17.5 kDa * Corresponding author. Fax: +1 765 494 0805.
E-mail address: [email protected] (M.T. Harris). 0021-9797/$ – see front matter © 2006 Published by Elsevier Inc. doi:10.1016/j.jcis.2005.12.068
wrapped in a helix around a single strand of RNA [17]. These particles are stable in solutions of approximately pH 2–10 and at temperatures up to 60 ◦ C [18]. The surface of the assembled virus particle contains numerous glutamic acid, aspartic acid, lysine and arginine residues, giving an overall negative surface charge above the isoelectric point, pH 3.0–3.5, or a positive charge when below. The length of the TMV particle is controlled by the length of the encapsulated viral RNA. However, a genetically modified TMV mutant, E50Q, contains a glutamic acid to glutamine mutation at position 50 within the coat protein open reading frame. This mutation neutralizes a negative–negative charge repulsion involved in the controlled assembly of the virion. As a result, the E50Q mutant produces virus particles that assemble beyond the length of the encapsulated RNA, averaging approximately 1000 nm in length yet still maintaining the properties of the wild-type virus [19]. Three pH domains for the polymerization of silica determine the morphology of the silicate oxide products: pH < 2, pH 2–7, and pH > 7 [20]. Previous examples of silica coatings created by Shenton et al. were conducted in strong acidic conditions (pH 2.5) where the acid catalyzed hydrolysis of TEOS in an ethanol solvent (between the isoelectric points of TMV (pI 3.5) [21] and silica (pI 2) [22]) produced a charge–charge
E. Royston et al. / Journal of Colloid and Interface Science 298 (2006) 706–712
interaction between the negatively charged silica and the positively charged TMV surface [10]. However, a search for milder reaction conditions in addition to a more detailed characterization of the silica-coated TMV is needed for the future application of TMV-based silicate oxide nanotubes. Milder reaction conditions at the slightly basic conditions are not only safer but provide greater control over gelation by slowing the overall gelation process in comparison to a similar reaction system under acidic conditions. At low ionic concentration (<2 mM monovalent ions) and low TEOS concentration, electrostatic repulsion plays a prominent role in slowing the gelation process above pH 8 for aqueous systems [20]. Furthermore, the rate of hydrolysis reduces while the rate of silica dissolution increases in higher pH conditions [22] resulting in an overall decrease in the rate of gelation. The research presented in this paper therefore achieves silica layer formation in aqueous methanolic solutions in a pH > 7. The hydrolysis and condensation reactions of TEOS in aqueous methanolic solutions produce denser nucleating particles about half the diameter of those formed under similar reaction conditions in ethanol [23]. These particles may help produce smoother coatings with increased density as a result of the increased mobility and smaller negatively-charged surface area of the smaller silica particles. Due to stability concerns with the virus in alcohol–water solutions, the concentration of water must be at least 40 wt%. Alcohol denatures proteins by disrupting the intramolecular hydrogen bonding, where new hydrogen bonds are formed instead between the alcohol molecule and the protein side chains [24]. Thus for purposes of this paper, concentrations of alcohol– water mixtures are maintained at 50–50 wt%. Characterization of the silica-TMV nanotubes is performed using electron energy loss spectroscopy (EELS), small angle X-ray scattering (SAXS), and dynamic light scattering (DLS). EELS provides high spatial resolution elemental analysis of the silica coating, while SAXS and DLS provide in situ information on the structures of the system. The detailed structural information resulting from this research will be useful in utilizing silica-TMV nanotubes in other applications, such as multi-layered coatings. 2. Experimental 2.1. Materials Tetraethylorthosilicate (Aldrich 98%) and ammonium hydroxide (NH4 OH, Aldrich 5N) were used as received. HPLC grade methanol (MeOH, 99%) and deionized water were used in all preparations. Both the E50Q mutant and wild-type TMV were prepared as previously described [19]. 2.2. Coating TMV A 50–50 wt% solution of alcohol to water was used as the solvent. The base catalyst was added to a total concentration of 0.001 M NH3 , followed by TMV to a total concentration of 0.2 mg/mL, and finally by TEOS to a total concentration of 0.009 M. The NH3 concentrations were kept low due to stability concerns with the TMV. The solution was shaken vigorously
707
between additions of each reactant. The reaction mixture was allowed to stand for 24 h to ensure successful growth of a silica shell around the virus. Reference buffer solutions for pH measurements in a 50–50 wt% methanol–water mixture were prepared as described by Paabo et al. [25]. The pH of the sample was measured at the start and end of the reaction and read at pH∗ = 8.8 for both reaction times. This result was expected as the change in pH due to the silica reaction should be small as the concentration of TEOS in solution is very low. To remove the excess silica from solution once the shell has been created, the sample was centrifuged at 12,500 rpm for 15 min several times and re-suspended in methanol. TEM images of the solution were taken after 24 h. 2.3. Characterization Transmission electron microscopy (TEM) was conducted using a Philips CM10 microscope. The particle size was verified from measurements taken from several negatives of each sample. All TEM samples were prepared without staining by placing a drop of sample on a carbon coated copper grid followed by wicking away the excess liquid. Electron energy loss spectroscopy (EELS) was carried out using a Vacuum Generator HB501 STEM. Cold-stage EELS studies were conducted using a JEOL 2010F TEM. The beam energies were 80 and 200 keV (100 and 200 keV max), respectively. Silica maps were taken on the 2010F using energy centers of 82, 92, and 113 eV with 10 eV spreads for 16 s, with the dried sample held at −174 ◦ C during the spectrum collection. Small angle X-ray scattering (SAXS) measurements were taken using a Molecular Metrology X-ray scattering instrument with a microfocus X-ray source, an Osmic MaxFlux confocal X-ray optic, a three-pinhole camera, and a gas-filled 2D multiwire detector at a 1424 mm camera length. Calibration in q-space was achieved with an isotropic silver behenate powder standard. The hydrodynamic radius of the particles was measured at room temperature using a dynamic light scattering (DLS) system (Photocor Instruments) that was equipped with a 20 mW He–Ne laser (λ = 632 nm). All measurements were taken at a 90◦ scattering angle. 3. Results and discussion TEM images of unstained silica-coated wild-type TMV and E50Q mutant TMV particles re-suspended in 100% methanol are shown in Figs. 1a and 1b, respectively. The inserts in Figs. 1a and 1b show uncoated and unstained wild-type TMV and E50Q in water. In Fig. 1a, the arrows indicate the length of one virion, since the image shows two virions lined end to end. The lengths of the nanotubes are consistent with those of wild-type TMV (300 ± 29 nm) and E50Q mutant TMV (940 ± 470 nm) templates [19]. Due to the low electron density of silica, the thin silica shell on the TMV is difficult to observe clearly. TEM observation at higher magnification re-
708
E. Royston et al. / Journal of Colloid and Interface Science 298 (2006) 706–712
(a)
(a)
(b)
Fig. 2. (a) DLS results for silica-coated wild-type TMV (1) in reaction conditions over time. (b) DLS results for silica-coated E50Q (P) in reaction conditions over time. The bars show one standard deviation, the points are the means. The shaded areas are the calculated Rh ranges.
(b) Fig. 1. (a) TEM image, unstained, of silica-coated wild-type TMV suspended in 100% MeOH. (b) TEM image, unstained, of silica-coated E50Q suspended in 100% MeOH. The inserts are unstained and uncoated wild-type TMV and E50Q respectively. In (a) the arrows indicate the length of one virion, since the image shows two virions lined end to end. Scale is the same between each insert and image.
sults in blurring of the image and failure to show a discrete silica shell. DLS studies help to show that the dispersion and stability of the virus particles remain unaffected by the coating process. The hydrodynamic radius of the silica-coated wild-type TMV before centrifugation and re-suspension in methanol is found to be around 55 nm, whereas the E50Q mutant has a hydrodynamic radius of around 60 nm, consistent with the presence of longer virion particles. Fig. 2 shows the Rh of the silicacoated TMV particles over time. From the theoretical diffusion coefficient for a cylinder [26,27] and the given values for the distribution of sizes for the uncoated virus, the approximate expected value of the Rh of wild-type TMV and E50Q falls between 50–60 nm and 70–160 nm, respectively. The experimental DLS values for Rh are in agreement with the theoretical values, as shown by the shaded areas in Fig. 2. Although there is significant overlap of the error bars, the mean Rh values for wtTMV and E50Q are statistically different at approximately the 90% confidence level. The smaller size range for the E50Q particles is likely a result of a combination of factors, such as fragmentation of the virus during the coating process due to the alcohol and ammonia in the solution and inherent length differences associated with different E50Q preparations. The increase in length of the silica-coated E50Q as compared to wtTMV agrees with TEM observations, as seen in Figs. 1a and 1b.
TMV relies on water bridging to maintain the conformational stability of its proteins [28]. Mixtures with concentrations higher than 60 wt% alcohol compromise the stability of the TMV structure. In these situations, the bridging water molecules are reoriented and displaced by alcohol molecules and the TMV disassembles. The results presented in this paper demonstrate stable silica-coated TMV particles in solutions of 100% methanol, indicating the presence of another “force” stabilizing the rod-like structure of the TMV. It is likely that the silica shell on the TMV surface prevents the reorientation of the water molecules, thus keeping the alcohol groups from binding with the protein. Re-suspension of the silica-coated TMV in water results in the breakup of the TMV structure over extended periods of time. This is most likely due to the increased solubility of silica in water versus the solubility of silica in methanol. Silica dissolves off the surface of the TMV and causes instability of the protein structure as it leaves. The exact nature of this instability has not been determined. TEM was used to monitor the effect of time on re-suspensions of silica-coated TMV in methanol, water, and silica-saturated water. After 1 h, silicacoated TMV in water shows definite signs of breakup, while the silica-coated TMV in methanol and silica-saturated water both contain stable virions. After 3 h, there is no sign of TMV in water, whereas unchanged stable virions are still seen in methanol and silica-saturated water solutions. The TEM results for samples suspended in methanol and silica-saturated water at 3 h and the result for a sample suspended in water at 1 h are shown in Fig. 3. No image is included for the silica-coated TMV re-suspended in pure water at 3 h as there are no TMV particles present on the grid. This result suggests that the association of silica with the TMV coat protein changes the nature of the molecular interaction between neighboring coat proteins in some fashion. Electron energy loss spectroscopy (EELS) and cold-stage EELS are used to confirm the presence and location of silica
E. Royston et al. / Journal of Colloid and Interface Science 298 (2006) 706–712
709
(a)
(a) (b)
(c) Fig. 3. TEM images of silica-coated TMV re-suspended in (a) MeOH after 3 h (b) silica-saturated H2 O after 3 h (c) H2 O after 1 h. An image of silica-coated TMV re-suspended in H2 O is not included as there are no particles present on the TEM grid. The scale bar is 300 nm in all images.
and TMV in the sample. EELS is chosen for its high spatial resolution (as low as 0.3–0.5 nm in some cases), unlike X-ray emission spectroscopy which suffers from beam broadening due to backscattering, secondary fluorescence, and the generation of fast secondary electrons in the specimen [29]. Scanning transmission electron microscopy (STEM) is used in conjunction with EELS, while TEM is used with the cold-stage EELS. Previous EELS studies conducted on TMV by Leapman et al. show the presence of carbon and phosphorous (from the RNA) indicate the existence of TMV [30]. Figs. 4a and 4b (EELS results) show the presence of phosphorous and additional carbon on the rods relative to the background, indicating the rod structures are TMV. The log-differential of the EELS signal intensity is plotted in order to show the peaks that are otherwise difficult to discern (Fig. 4b). To verify the presence of the silica on the TMV, a silica map of the silica-coated TMV sample is made. Cold-stage EELS is used in order to minimize the effects of beam damage present
(b) Fig. 4. (a) Darkfield TEM image of silica-coated wild-type TMV. Silica appears in the background as bright clusters. (b) Log-differential plot of the EELS data displaying the peaks with clarity. The data corresponds to the areas indicated by the arrows in (a).
in a regular EELS sample holder. The silica map of the silicacoated wild-type TMV re-suspended in methanol is seen in Fig. 5b. The average diameter of the TMV is 19 ± 0.5 nm. Results in Figs. 5a and 5b show an evenly distributed silica presence over the entire surface of the TMV, which would not be the case if this was driven by capillary forces during drying. In that case, the silica would demonstrate a higher presence along the outside edges of the TMV image, creating an outline of electron dense material. SAXS studies clearly show the difference between the silicacoated TMV and uncoated TMV particles in the presence of preformed silica particles. The preformed silica particles are made using the same reaction condition as that used for the silica-coated TMV. The S/N ratio of the instrument is too low
710
E. Royston et al. / Journal of Colloid and Interface Science 298 (2006) 706–712
(a)
(a)
(b) Fig. 6. SAXS data demonstrating the differences between (1) the reaction of TEOS in the presence of TMV, (P) the reaction of TEOS without TMV, (E) TMV in water, and (!) TMV added to TEOS reacted without TMV present. (a) Intensity vs Q plot of the data. (b) ln(I ) vs Q2 plot showing linearity of data for Guinier analysis. (b) Fig. 5. (a) TEM image of silica coated wild-type TMV. (b) Silica map taken using cold-stage EELS (−174 ◦ C).
at the lower intensity levels to clearly show the expected smaller peaks associated with TMV [31]; therefore, the SAXS results, as shown in Fig. 6, are used to demonstrate the differences between the systems, in situ. Studies on the length of exposure to the beam and the degradation of the sample indicate no measurable damage during the 1 h sample collection time. Fig. 6a shows the scattering intensity versus Q (nm−1 ) data. The SAXS results show a decrease in scattering in the 0.2– 1 nm−1 Q range when the TEOS is reacted to form silica in the presence of TMV versus scattering when the silica is preformed and is subsequently mixed with TMV. The scattering curve for the sample with preformed silica particles and uncoated TMV is a combination of the scattering from the silica particles and the TMV in water. The decreased scattering intensity for the silica-coated TMV sample may be due to a continuous coating of small silica structures on the TMV surface. USAXS experiments are planned to determine if more intense scattering occurs at lower Q for the silica-coated TMV system,
Table 1 Rg values calculated from SAXS data over 0.5 < Q2 < 2.25
Silica-coated TMV Supernatant Concentrated re-suspension Silica particles Silica + TMV
Rg
R2
1.70 1.84 0.86 1.86 1.92
0.989 0.992 0.893 0.993 0.998
which would indicate the formation of a continuous coating of silica on the TMV particles and would account for the decrease in scattering intensity, I , in the ∼0.2 nm−1 Q range. ln(I ) versus Q2 plots are shown in Fig. 6b to determine if there is a sufficient Guinier region to obtain radius of gyration, Rg , values. The scattering intensity data in the region 0.5 < Q2 < 2.25 show a linear behavior. The results of the linear regression analysis are given in Table 1 and are discussed below. Fig. 7 shows the SAXS results detailing the effectiveness of centrifugation in separating the unattached silica particles from the silica-coated TMV particles. This step is essential in minimizing residual silica particles during the preparation of
E. Royston et al. / Journal of Colloid and Interface Science 298 (2006) 706–712
(a)
(b) Fig. 7. (a) SAXS data of (1) the reaction of TEOS in the presence of TMV, (P) TMV in water, (E) the supernatant from the separation of silica-coated TMV from the silica background, and (!) re-suspension of silica-coated TMV in MeOH. (b) ln(I ) vs Q2 plot showing linearity of data for Guinier analysis for (1) the reaction of TEOS in the presence of TMV, (E) the supernatant from the separation of silica-coated TMV from the silica background, (!) re-suspension of silica-coated TMV in MeOH, and (e) the concentrated re-suspension of the silica-coated TMV in MeOH.
silica-coated TMV samples for all TEM images. It is also important to remove the residual silica for the EELS analysis. The majority of the silica and silica-coated TMV is found in the supernatant since the scattering curves (Figs. 7a and 7b) for the silica-coated TMV sample and the supernatant almost overlap. The re-suspended pellet shows a scattering curve (Fig. 7a) containing TMV without background silica. Unfortunately, the X-ray scattering curves display a very noisy pattern (Fig. 7b) when the suspension is diluted to the original sample volume and it renders Guinier analysis impossible. Therefore, the suspension is concentrated by a factor of six over the course of the multiple centrifugations. The SAXS data for the concentrated suspension is also shown in Fig. 7b and is used in the Guinier analysis for silica-coated TMV re-suspended in MeOH. The radius of gyration (Rg ) values calculated from the linear section of the ln(I ) vs Q2 curve (slope = Rg2 /3), in addition to the R 2 values for the linear fit, are reported in Table 1. From the results, it appears that the Rg values (approximately 1.8–1.9 nm) are for the silica particles in solution. The radii of gyration for these particles are approximately half the size of the first nanophases
711
that are reported by Green et al. [23] for the hydrolysis and condensation of TEOS in aqueous-methanolic solutions and at much lower water concentrations. The decrease in the size of silica nanophase that is observed in the current experiments is most probably due to the higher water concentration (approximately 24 mol/L) during the hydrolysis and condensation of TEOS. Aggregates of silica particles are confirmed in the background of the STEM image, Fig. 4a, with the smaller particles being on the order of 2 nm in diameter. As expected, the Rg values for the silica (1.86 nm) and the silica + TMV (1.92 nm) systems are similar since, relative to the scattering from the silica particles, the scattering by TMV is very low in the Q range where the Guinier analysis is done. In other words, scattering by the silica particles dominates in this Q range. The silica nanostructures in the supernatant have an Rg value of 1.84 nm which is similar to that of the silica particle system. Thus, 1.8–1.9 nm silica nanostructures are still being formed when TEOS is reacted to form silica in the presence of TMV; however, the lower intensity of the scattering curve relative to the scattering curve for the silica + TMV system in the 0.2–1.5 nm Q range indicates that fewer of these nanostructures are present. The Rg (0.9 nm) associated with the nanostructures in the concentrated re-suspension of the silica-coated TMV are possibly due to the presence of smaller nanostructures that detach from the surface of the silica-coated TMV when the methanol is added to re-suspend the sample. The mechanism for the coating of silica on TMV by the hydrolysis and condensation of TEOS is not understood. Hydrolyzed and partially condensed species of TEOS may be attracted to the surface of the TMV due to attractive particle interactions like van der Waals forces. In addition, the negatively charged silica monomers may be attracted to the positively charged arginine and lysine amino acids on the surface of the TMV. Furthermore, studies investigating the interaction mechanism between proteins and hydrophilic silica surfaces [32] indicate that interactions between hydroxyl groups on the silica surface and carboxyl/carbonyl or amine/imido groups [33] in the protein structure should contribute to the formation of the silica shell on the TMV surface. Once the initial hydrolysis is completed and the partially condensed species are attached to the surface of the TMV, additional silica precursors may participate in condensation reactions with the attached silicates to form a gel on the surface of the TMV. Additional research is required to fully elucidate the reaction. 4. Conclusion Silica-coated TMV has been successfully produced by the reaction of TEOS in methanol and water at pH values above those that have been previously reported. Stabilization of the TMV in methanol has been achieved through the silica coating. TEM studies combined with EELS analysis confirm the presence of silica on the surface of the TMV while SAXS studies show the differences between uncoated and coated TMV particles. The E50Q mutant has also been silica-coated successfully, giving rise to longer silica nanostructures. Successful charac-
712
E. Royston et al. / Journal of Colloid and Interface Science 298 (2006) 706–712
terization of the system provides the information required to enhance the usefulness of the silica-coated TMV as a template. Acknowledgments This research is supported by the U.S. Department of Energy (DEFG02-02-ER45975 and DEFG02-02-ER45976). TEM work is done in the Life Science Microscopy Facility, Purdue University. EELS analysis is carried out in the Center for Microanalysis of Materials, University of Illinois, which is partially supported by the U.S. Department of Energy under Grant DEFG02-91-ER45439. Special thanks go to Dr. Ray D. Twesten for his assistance in collecting the EELS data. SAXS studies are supported by the NSF funded (MRI program award 0321118-CTS) facility for in situ X-ray scattering from nanomaterials and catalysts, Purdue University. References [1] A. Huezko, Appl. Phys. A: Mater. Sci. Process. 70 (2000) 365. [2] X. Gao, H. Matsui, Adv. Mater. 17 (2005) 2037. [3] R. Kirsch, M. Mertig, W. Pompe, R. Wahl, G. Sadowski, K.-J. Böhm, E. Unger, Thin Solid Films 305 (1997) 248. [4] J.L. Coffer, S.R. Bigham, X. Li, R.F. Pinizzotto, Y. Rho, G. Young, R.M. Pirtle, I.L. Pirtle, Appl. Phys. Lett. 69 (1996) 3851. [5] M. Markowitz, S. Baral, S. Brandow, A. Singh, Thin Solid Films 224 (1993) 242. [6] T. Douglas, M. Young, Adv. Mater. 11 (1999) 679. [7] Y. Huang, C.Y. Chiang, S.K. Lee, Y. Gao, E.L. Hu, J. De Yoreo, A.M. Belcher, Nano Lett. 5 (2005) 1429. [8] C.E. Flynn, S.-W. Lee, B.R. Peelle, A.M. Belcher, Acta Mater. 51 (2003) 5867. [9] C. Mao, D.J. Solis, B.D. Reiss, S.T. Kottman, R.Y. Sweeney, A. Hayhurst, G. Georgiou, B. Iverson, A.M. Belcher, Science 303 (2004) 213.
[10] W. Shenton, T. Douglas, M. Young, G. Stubbs, S. Mann, Adv. Mater. 11 (1999) 253. [11] E. Dujardin, C. Peet, G. Stubbs, J.N. Culver, S. Mann, Nano Lett. 3 (2003) 413. [12] W.L. Liu, K. Alim, A.A. Balandin, Appl. Phys. Lett. 86 (2005) 253108. [13] S. Lee, E. Royston, J.N. Culver, M.T. Harris, Nanotechnology 16 (2005) S435. [14] M. Knez, M. Sumser, A.M. Bittner, C. Wege, H. Jeske, S. Kooi, M. Burghard, K. Kern, J. Electroanal. Chem. 522 (2002) 70. [15] M. Knez, A.M. Bittner, F. Boes, C. Wege, H. Jeske, E. Maiß, K. Kern, Nano Lett. 3 (2003) 1079. [16] C.E. Fowler, W. Shenton, G. Stubbs, S. Mann, Adv. Mater. 13 (2001) 1266. [17] K. Namba, R. Pattanayek, G. Stubbs, J. Mol. Biol. 208 (1989) 307. [18] G. Stubbs, Seminar Virology 1 (1990) 405. [19] B. Lu, G. Stubbs, J.N. Culver, Virology 225 (1996) 11. [20] R.K. Iler, The Chemistry of Silica, Wiley, New York, 1979. [21] G. Oster, J. Biol. Chem. 190 (1951) 55. [22] C.J. Brinker, G.W. Scherer, Sol–Gel Science, Academic Press, New York, 1990. [23] D.L. Green, J.S. Lin, Y. Lam, M.Z.-C. Hu, D.W. Schaefer, M.T. Harris, J. Colloid Interface Sci. 266 (2003) 346. [24] A.L. Lehninger, D.L. Nelson, M.M. Cox, Principles of Biochemistry, second ed., Worth Publishers Inc., New York, 1993. [25] M. Paabo, R.A. Robinson, R.G. Bates, J. Am. Chem. Soc. 87 (1965) 415. [26] M. Doi, S.F. Edwards, The Theory of Polymer Dynamics, Oxford Univ. Press, Oxford, 1986. [27] N.C. Santos, M.A.R.B. Castanho, Biophys. J. 71 (1996) 1641. [28] A.C. Bloomer, J.N. Champness, G. Bricogne, R. Staden, A. Klug, Nature 276 (1978) 362. [29] R. Egerton, Electron Energy-Loss Spectroscopy in the Electron Microscope, second ed., Plenum Press, New York, 1996. [30] R.D. Leapman, N.W. Rizzo, Ultramicroscopy 78 (1999) 251. [31] A.G. Malmon, Acta Crystallogr. 10 (1957) 639. [32] Y.I. Tarasevich, L.I. Monakhova, Colloid J. 64 (2002) 482. [33] D. Christendat, T. Abraham, Z. Xu, J. Masliyah, J. Adhes. Sci. Technol. 19 (2005) 149.