Dramatic Thermal Stability of Virus−Polymer Conjugates in

pubs.acs.org/Langmuir © 2010 American Chemical Society

Dramatic Thermal Stability of Virus-Polymer Conjugates in Hydrophobic Solvents Patrick G. Holder,† Daniel T. Finley,† Nicholas Stephanopoulos,† Ross Walton,^ Douglas S. Clark,‡ and Matthew B. Francis*,†,§ †

Department of Chemistry, ‡Department of Chemical and Biomolecular Engineering, University of California, Berkeley, California 94720, United States, §Materials Sciences Division, and ^Life Science Division, Lawrence Berkeley National Laboratory, Berkeley, California 94720, United States Received September 30, 2010

We have developed a method for integrating the self-assembling tobacco mosaic virus capsid into hydrophobic solvents and hydrophobic polymers. The capsid was modified at tyrosine residues to display an array of linear poly(ethylene glycol) chains, allowing it to be transferred into chloroform. In a subsequent step, the capsids could be transferred to a variety of hydrophobic solvents, including benzyl alcohol, o-dichlorobenzene, and diglyme. The thermal stability of the material against denaturation increased from 70 C in water to at least 160 C in hydrophobic solvents. With a view toward material fabrication, the polymer-coated TMV rods were also incorporated into solid polystyrene and thermally cast at 110 C. Overall, this process significantly expands the range of processing conditions for TMV-based materials, with the goal of incorporating these templated nanoscale systems into conductive polymer matrices.

Introduction The development of materials structured by viral capsids is progressing rapidly, based on the observation that self-assembling protein templates can display periodic functionality in a variety of geometric orientations.1-3 The tobacco mosaic virus (TMV) serves as a particularly successful example of this strategy by providing a rodlike template for inorganic materials,4-7 arrays of porphyrins,8 and platinum nanowires.9 In terms of higher-order assemblies, it was aligned into elongated fibers using shear force.10 In our lab, TMV is used as a structural scaffold for displaying affinity tags,11 arraying light-harvesting chromophores,12 and aligning carbon nanotubes.13 An important next step forward in the generation of viral capsid-based electronic materials is the integration of functionalized capsids with conductive materials. Indeed, interesting recent work shows that this goal can be achieved with water-soluble polyaniline and polypyrrole polymers.14 However, it is often a difficult task to integrate viral capsids with, for example, the hydrophobic hole-conducting polymers polythiophene or *Corrresponding author. E-mail: [email protected].

(1) Douglas, T.; Young, M. Nature 1998, 393, 152–155. (2) Wang, Q.; Lin, T.; Tang, L.; Johnson, J. E.; Finn, M. G. Angew. Chem., Int. Ed. 2002, 41, 459–462. (3) Young, M.; Debbie, W.; Uchida, M.; Douglas, T. Annu. Rev. Phytopathol. 2008, 46, 361–384. (4) Shenton, W.; Douglas, T.; Young, M.; Stubbs, G.; Mann, S. Adv. Mater. 1999, 11, 253–256. (5) Knez, M.; Bittner, A. M.; Boes, F.; Wege, C.; Jeske, H.; Mai, E.; Kern, K. Nano Lett. 2003, 3, 1079–1082. (6) Dujardin, E.; Peet, C.; Stubbs, G.; Culver, J.; Mann, S. Nano Lett. 2003, 3, 413–417. (7) Douglas, T.; Young, M. Adv. Mater. 1999, 11, 679–681. (8) Endo, M.; Fujitsuka, M.; Majima, T. Chem.;Eur. J. 2007, 13, 8660–8666. (9) Liu, W. L.; Alim, K.; Balandin, A. A.; Mathews, D. M.; Dodds, J. A. Appl. Phys. Lett. 2005, 86, 253108–3. (10) Kuncicky, D. M.; Naik, R.; Velev, O. Small. 2006, 2, 1462–1466. (11) Schlick, T.; Ding, Z.; Kovacs, E.; Francis, M. J. Am. Chem. Soc. 2005, 127, 3718–3723. (12) Miller, R.; Presley, A.; Francis, M. J. Am. Chem. Soc. 2007, 129, 3104–3109. (13) Holder, P. G.; Francis, M. B. Angew. Chem., Int. Ed. 2007, 46, 4370–4373. (14) Niu, Z.; Liu, J.; Lee, L.; Bruckman, M.; Zhao, D.; Koley, G.; Wang, Q. Nano Lett. 2007, 7, 3729–3733.

Langmuir 2010, 26(22), 17383–17388

poly(phenylenevinylene). In a general sense, this challenge arises because multimeric protein-based assemblies cannot be readily dissolved in organic solvents or exposed to the high temperatures often required for lithography and thermal annealing. The ability to use such processing would be highly advantageous for expanding the number and type of functional optoelectronic components that can be integrated with the exquisite periodic arrays afforded by virus-based materials. With these goals in mind, we set out to address and understand the inherent incompatibility of hydrophilic protein surfaces with hydrophobic solvents. When a protein is introduced into mixtures of water and cosolvents;for the purpose of solubilizing exogenous functional components;the thermodynamically most stable conformation of the protein would be expected to be different from that in aqueous media alone. If a kinetic pathway exists for the protein backbone to reorganize, this can at a minimum disassemble capsid quaternary structure and at worst denature monomer proteins. A more successful strategy would involve the transfer of proteins into nonpolar organic solvents that lack hydrogen-bond-donating and -accepting capabilities, as this should kinetically trap the electrostatic interactions that hold the secondary and tertiary structural elements together. In this environment, the protein should be more likely to remain in its native state even if more stable conformations exist. Past work shows that transferring capsids into organic solvents can be achieved using ionic surfactants that pair with charged groups on the protein surface15-17 (these have, in our hands, failed for TMV). Similarly, ferritin cages were transferred into dichloromethane solutions through the attachment of hydrophobic chains18,19 and (15) Johnson, H. R.; Hooker, J. M.; Francis, M. B.; Clark, D. S. Biotechnol. Bioeng. 2007, 97, 224–34. (16) Moriyama, Y.; Takeda, K. Langmuir. 2005, 21, 5524–5528. (17) Brochette, P.; Petit, C.; Pileni, M. P. J. Phys. Chem. 1988, 92, 3505–3511. (18) Wong, K. K. W.; Whilton, N. T.; Douglas, T.; Mann, S.; Colfen, H. Chem. Commun. 1998, 1621–1622. (19) Wong, K. K. W.; C€olfen, H.; Whilton, N. T.; Douglas, T.; Mann, S. J. Inorg. Biochem. 1999, 76, 187–195.

Published on Web 10/21/2010

DOI: 10.1021/la1039305

17383

Article

Holder et al.

polymers.20-22 In our lab, we have shown that TMV capsids displaying PEG chains are sufficiently hydrophobic to be transferred from water to chloroform.11 All of these systems provide interesting opportunities to study the effects of solvent on capsid stability and potentially access an expanded range of material processing conditions. In this report, we expand our initial results with TMV to include a number of different solvents that can be used for thin film casting, with the observation that the structures can withstand very high temperatures (at least up to 160 C). We also demonstrate that modified TMV capsids can be dissolved in polymerizable monomers, subsequently affording capsid-containing plastics that can be cast thermally into different shapes, without degradation of the internal self-assembled protein structures.

Scheme 1. Synthesis of the Virus-Polymer Hybrid Rods and Their Transfer into Organic Solvent

Results and Discussion The native TMV capsid is a 300 nm long cylinder that is 18 nm in diameter. It can be harvested on the gram scale from plant infection,11 or it can be obtained without length control through heterologous expression of the coat protein in E. coli.12 The capsid presents four potential surfaces for chemical modification: the 50 and 30 “ends” of the cylinder (named for the corresponding termini of its encased RNA chain) as well as the interior and exterior walls of the tube.11 These latter two surfaces present an array of more than 2000 copies of the solution-accessible amino acids encoded into the TMV monomer. Synthesis and Purification of Virus-Polymer Conjugates. In order to solubilize TMV in organic solvents, we turned to proteinpolymer conjugates. These materials can be used to incorporate hydrophilic proteins with functional polymeric components in biohybrid surfactants,23 as telechelic conjugates,24-26 and for the selective sequestration of toxic ions.27 For our purposes, we chose linear poly(ethylene glycol) (PEG) as a protective coating for the viral material because of its compatibility with both aqueous and organic solvents. We have shown previously that these polymer chains can be attached by first targeting an exposed tyrosine residue (Y139) on each of the protein monomers with a diazonium salt (Scheme 1).11 These previous studies have also confirmed the site selectivity through the analysis of modified TMV protein through proteolytic digest followed by mass spectrometry. The uniquely reactive ketone group installed using this reaction can then be modified using alkoxyamines in a subsequent step. Using a 2 kDa poly(ethylene glycol) alkoxyamine,11 oxime formation proceeded under aqueous conditions (pH 6.0) in an overnight reaction. Notably, not all of the ketone sites on the TMV rod were modified with PEG chains when reacted with a range of 1-50 mol equiv of PEG, as shown by gel electrophoresis (see Supporting Information Figure S2). This should allow for further modification of unreacted ketones with additional functionality in the future. Upon addition of 50 equiv of PEG, densitometry analysis after Coomassie staining indicated that (20) Zeng, Q.; Li, T.; Cash, B.; Li, S.; Xie, F.; Wang, Q. Chem. Commun. 2007, 1453–1455. (21) Sengonul, M.; Ruzicka, J.; Attygalle, A. B.; Libera, M. Polymer 2007, 48, 3632–3640. (22) Escosura, A. D. L.; Nolte, R. J. M.; Cornelissen, J. J. L. M. J. Mater. Chem. 2009, 19, 2274–2278. (23) Reynhout, I. C.; Cornelissen, J. J. L. M.; Nolte, R. J. M. Acc. Chem. Res. 2009, 42, 681–692. (24) Bays, E.; Tao, L.; Chang, C.; Maynard, H. D. Biomacromolecules 2009, 10, 1777–1781. (25) Tolstyka, Z. P.; Kopping, J. T.; Maynard, H. D. Macromolecules 2008, 41, 599–606. (26) Boyer, C.; Bulmus, V.; Davis, T. P.; Ladmiral, V.; Liu, J.; Perrier, S. Chem. Rev. 2009, 109, 5402–5436. (27) Esser-Kahn, A. P.; Iavarone, A. T.; Francis, M. B. J. Am. Chem. Soc. 2008, 130, 15820–15822.

17384 DOI: 10.1021/la1039305

a Two iterations of N-hydroxyphthalimide, DIAD, PPh3, CH2Cl2. Aqueous H2NNH2, CH2Cl2. c p-Aminobenzophenone, NaNO2, tosylic acid, borate buffer pH 9. d 2, phosphate buffer, pH 6.0. e Precipitate with 1.9 M (NH4)2SO4, centrifugation. f Dissolve in CHCl3, dry over Na2SO4, filter. b

50% of the available sites displayed a PEG chain. Here, 50% modification corresponds to ∼1000 PEG chains or at least 2 MDa of polymer per capsid. Before transferring them to organic solvents, the capsids were subjected to multiple rounds of centrifugal concentration against 100 kDa MWCO filters to remove small molecules and some of the unconjugated PEG chains. Virus-Polymer Conjugate Materials in Hydrophobic Organic Solvents. PEG-TMV (6) was transferred into an organic solvent by taking advantage of the property of PEG to precipitate from aqueous solutions at high ionic strength. Since the protein was covalently encased within a PEG shell, it also precipitated upon addition of 1.9 M (NH4)2SO4; in contrast, ketone-bearing TMV rods lacking the PEG chains remained in solution under these conditions. Upon replacing the water with chloroform, the material redissolved. The organic solution was then treated with a drying agent (Na2SO4) in order to remove any water that remained associated with the PEG chains. After filtering to remove the drying agent, the PEG-TMV was stored in the chloroform solution. We have stored virus materials in this condition at 4 C for periods up to 5 years with no apparent degradation. The resulting hybrid material was then characterized by atomic force microscopy (AFM) using low force constant silicon cantilevers, which minimally perturb the soft protein material in order to measure accurate height dimensions (Figure 1). AFM images taken after spin-coating 6 from chloroform solution confirmed that the viral capsids remained intact and that they retained their high aspect ratio, measuring ∼20 nm wide and ∼200 nm long. Phase-mode images also indicated a large halo of material surrounding the virus to be a different material than the protein, presumably representing both the covalently attached PEG layer and associated polymer chains that were not removed by the centrifugal concentration steps. Langmuir 2010, 26(22), 17383–17388

Holder et al.

Figure 1. AFM characterization of 6 deposited from chloroform solution. (a-c) were purified by centrifugal ultrafiltration, and (d) was purified by gel filtration as well as ultrafiltration. (a) Height image of two virus rods. (b) Phase image of the rods showing the three different materials: mica surface, polymer (dark), and protein (bright). (c) Sectional measurements taken from the height image show that each rod is ∼16-17 nm tall, 175-200 nm long, and ∼20 nm wide. (d) Image of the FPLC-purified TMV rods demonstrates removal of the excess PEG from solution.

After seeing the large volume of material surrounding the TMV capsids in the AFM, and having already dehydrated the solution with a desiccant, we suspected there might be a significant amount of unbound polymer surrounding each capsid. We considered that the extra PEG chains in solution might influence the solubility of the virus material in organic solvents and/or unduly stabilize the conjugated material. To remove the excess material in subsequent preparations, the aqueous samples of polymerconjugated TMV capsids were purified by size exclusion chromatography (Sephacryl S-200 resin) before transfer to the organic solvents. As shown in Figure 1d, AFM imaging of a sample of PEG-TMV cast from chloroform solution after this step did not show the same halo around each virus particle. Much like samples of unmodified TMV cast from aqueous solution, minimal aggregation was observed in these images, suggesting that the individual capsids were well-dispersed by their polymer coating. Close inspection of the AFM in Figure 1d indicates the lengths of the TMV rods are shorter than 300 nm and vary from rod to rod. This is a result of the purification process, which frequently leads to capsid fragmentation during centrifugal concentration. Efforts are underway to optimize the purification process for the production of full-length rods. One might also expect the persistence length of the capsids to be significantly different as the solvent is changed from aqueous to organic. These effects remain unexplored and will be the subject of future experiments. The current data show, however, that the overall structure of the capsid assembly remains intact after excess polymer removal and organic solvent transfer has taken place. These virus-polymer materials could be readily transferred into other less polar organic solvents by adding the new solvent Langmuir 2010, 26(22), 17383–17388

Article

and evaporating the lower boiling chloroform under reduced pressure. In this way, we transferred PEG-TMV into o-dichlorobenzene, benzyl alcohol, diglyme (2-methoxyethyl ether), nitrobenzene, m-xylenes, ethylbenzene, and styrene. The use of very polar and efficient hydrogen-bond-donating solvents (such as methanol or acetonitrile) led to immediate precipitation of the protein, presumably after disassembly and denaturation occurred. Stability of PEG-TMV at Elevated Temperature. Although one might expect the most thermodynamically stable folding state of proteins to be quite different in hydrophobic environments, the hydrogen-bonding networks that stabilize secondary and tertiary structure are kinetically trapped due to the absence of donating and accepting groups from the solvent.28 Transfer of proteins to organic solvents, in either insoluble29,30 or solubilized forms,31,32 rigidifies the tertiary structure and decreases the rate of protein motions. As a result, several groups have observed improved stability for enzymes solubilized in isooctane using ionic surfactants,15-17 and a number of studies show that enzymes can be incorporated into organic solvents and polymers while retaining their activity.33-35 Studies also show that multimeric enzymes can retain their quaternary structure in the presence of sodium dodecyl sulfate (SDS) due to intrinsically high kinetic stability.36 This phenomenon was demonstrated for MS2 capsids transferred into isooctane using aerosol-OT as an anionic surfactant.15 To determine whether the hydrophobic environment improved the stability of the modified TMV assemblies at increased temperatures, we used dynamic light scattering (DLS) measurements as a convenient means to characterize the PEG-TMV samples before and after heating in a series of hydrophobic solvents (Figure 2). It is important to note that the nonspherical nature of the TMV capsids render these data approximate, as do the estimates of the widely varying solvent refractive indices. Nonetheless, these measurements provide a straightforward way to determine whether the protein has remained in the native assembled state, dissociated into monomeric species, or formed large denatured coagulates. Our purifications yielded TMV samples with an average length of 294 nm, as measured in aqueous buffer by DLS.13 Following modification with PEG and purification by centrifugal concentration, the polymer-TMV conjugates measured 318 nm in the same buffer and contain some fragmented rods that appear as a shoulder in the DLS analyses. When transferred to organic solvents, we noted an overlapping range of particle size distributions, with peak maxima appearing between 229 and 274 nm (Figure 2). After heating to 160 C in BnOH for 15 min, DLS of the cooled solution showed a similar size distribution, with an apparent maximum of 309 nm. It is important to note that for each thermal experiment the samples were cooled to room temperature but were not refiltered before measuring their average size. Similar results were obtained after heating in diglyme, with a maximum at 385 nm, and in o-dichlorobenzene, with a maximum at 232 nm (Table 1). Two of these solvents (o-dichlorobenzene and diglyme) demonstrated a (28) Rodriguez-Larrea, D.; Minning, S.; Borchert, T. V.; Sanchez-Ruiz, J. M. J. Mol. Biol. 2006, 360, 715–724. (29) Affleck, R.; Haynes, C. A.; Clark, D. S. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 5167. (30) Eppler, R. K.; Hudson, E. P.; Chase, S. D.; Dordick, J. S.; Reimer, J. A.; Clark, D. S. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 15672–15677. (31) Wangikar, P. P.; Michels, P. C.; Clark, D. S.; Dordick, J. S. J. Am. Chem. Soc. 1997, 119, 70–76. (32) Hudson, E. P.; Eppler, R. K.; Beaudoin, J. M.; Dordick, J. S.; Reimer, J. A.; Clark, D. S. J. Am. Chem. Soc. 2009, 131, 4294–4300. (33) Kim, J.; Kosto, T. J.; Manimala, J. C.; Nauman, E. B.; Dordick, J. S. AIChE J. 2001, 47, 240–244. (34) Novick, S. J.; Dordick, J. S. Biomaterials 2002, 23, 441–448. (35) Akbar, U.; Aschenbrenner, C. D.; Harper, M. R.; Johnson, H. R.; Dordick, J. S.; Clark, D. S. Biotechnol. Bioeng. 2007, 96, 1030–1039. (36) Manning, M.; Colon, W. Biochemistry 2004, 43, 11248–11254.

DOI: 10.1021/la1039305

17385

Article

Holder et al.

Figure 2. Stability of 6 against hydrophobic organic solvents and heat. (a) Conditions for transfer to other solvents: i, add highboiling solvent; ii, remove CHCl3 at reduced pressure. (b) Quantification of length distributions by dynamic light scattering. Peaks smaller than 300 nm are rod fragments that result from centrifugal concentration. Peaks larger than 300 nm are aggregates of multiple rods. Table 1. Size and Polydispersity of 6 in Various Solvents, with and without Heating for 20 mina solvent

peak 1

peak 2

PDI

318 -f 0.33 100 mM KPi, pH 7.4b 283 89 0.55 chloroformb b 274 150 0.53 diglyme d 385 0.21 diglyme þ Δ 240 0.28 benzyl alcoholb 309 0.40 benzyl alcohol þ Δe b 240 48 0.97 o-dichlorobenzene c f 232 0.64 o-dichlorobenzene þ Δ 229 0.39 nitrobenzeneb 554 0.72 nitrobenzene þ Δe b 1140 0.58 m-xylenes a Measurements were taken at 25 C. b Not heated. c After 120 C. d After 125 C. e After 160 C. f Shoulder not sufficiently resolved.

distribution of capsids among individual capsids and fragments of individual capsids after transfer at room temperature, as evidenced by a bimodal DLS spectrum. However, after heating, the capsids redistributed in solution into a single DLS peak representing an average aggregate size. Nitrobenzene and m-xylenes demonstrated enhanced aggregation, as evidenced by 500-1000 nm species in solution. Although the distribution maxima vary depending on the propensity of 6 to aggregate in each solvent, the high degree of overlap between the curve areas suggests that minimal structural changes have occurred. Importantly, these analyses did not show either individual proteins dissociated from the virus during the heating process (3-9 nm), nor did they indicate denatured coagulates (5þ μm), as can be observed in a control experiment in which the protein was intentionally denatured (see Supporting Information Figure S4). 17386 DOI: 10.1021/la1039305

Figure 3. Imaging the stability of 6 against organic solvents and heat. (a) AFM of the PEG-TMV in CHCl3 after heating to 57 C for 18 h. (b) Sectional measurements of the image in (a) confirm the capsids are intact. Each rod retains a height (diameter) of ∼20 nm and lengths from 50 to 200 nm. Unstained TEM images of the rods deposited from diglyme before (c) and after (d) heating for 20 min at 125 C as well as TEM images of PEG-TMV deposited from CHCl3 and stained with UO2(OAc)2 after heating for 20 min to 160 C in BnOH (e) and o-DCB (f) demonstrate the integrity of the rod structure.

To confirm the changes occurring in the bulk solution qualitatively, we imaged samples of PEG-TMV before and after heating in a variety of solvents. Figure 3a,b presents AFM characterization of a sample of PEG-TMV in chloroform after heating to 57 C for 18 h. Importantly, the rods remain intact, as evidenced by the unchanged dimensions of the TMV rods. We also characterized TMV samples that were heated to high temperature in hydrophobic organic solvents by transmission electron microscopy (TEM). Figure 3c,d shows unstained TEM images of PEG-TMV samples deposited on a holey carbon grid before and after being heated to 120 C for 20 min in diglyme. Intact rods in clusters were observed, which were similar to images obtained for samples cast from an aqueous solution that was not heated. Importantly, the TMV features commonly visualized by TEM are still apparent, including its dimensions and its central channel. Similar results were obtained after heating PEGTMV in benzyl alcohol and o-dichlorobenzene (Figure 3, parts e and f, respectively). Taken together, we interpret these data to show that the polymer-TMV conjugates do not undergo noticeable changes when transferred to nonpolar organic solvents. Further, the behavior of individual particles in solution is governed by the PEG coating on the capsid. Most notably, the stability of the capsids against high temperatures is unprecedented for protein assemblies. Incorporation of PEG-TMV into Hydrophobic Polymers. As a test of the compatibility of PEG-TMV with hydrophobic polymers, we incorporated PEG-TMV into solid polystyrene. Neat styrene was added to the PEG-TMV in CHCl3, and the Langmuir 2010, 26(22), 17383–17388

Holder et al.

Article

polymerization conditions, we also heated the PEG-TMV material in ethylbenzene for 12 h at 60 C. AFM images confirmed the expected features for PEG-TMV after heating (Figure 4e,f).

Conclusions These experiments show that TMV can be integrated with hydrophobic materials by surrounding the quaternary protein structure with poly(ethylene glycol) chains. This method of passivation allows dual compatibility with aqueous and organic solvents, and the modification leaves open sites on the virus surface for further elaboration with the growing library of aminooxy-containing functional components. We envision that this type of modification could be expanded to encompass other viral scaffolds that present a regular array of sites for modification. In addition, these materials demonstrate a highly useful tolerance against thermal denaturation in organic solvents, highlighting that their robust selfassembling structure is further enhanced by stabilization of the electrostatic networks along the protein surface. It is encouraging that the TMV-containing materials can be processed at elevated temperature into solid polystyrene. We are optimistic that the PEG-TMV structure will tolerate processing with other, more electronically interesting polymers and materials.

Materials and Methods

Figure 4. Synthesis, casting, and characterization of polystyrene embedded with PEG-TMV. (a) The chloroform solution was diluted with neat styrene, and then the chloroform was removed under reduced pressure. Polymerization was induced with an initiator (AIBN) and was allowed to proceed at 60 C for 12 h. (b) Photograph of the plastic embedded with the hybrid material as cast in a glass vial (top left) and a thermal casting taken of the Roosevelt dime (bottom left). The dime itself is shown for scale and feature comparison. (c) TEM image, negative stained with 2% w/v aqueous UO2(OAc)2, of TMV rods reisolated by dissolving a small fragment of the polystyrene in CHCl3. (d) Control experiment in which PEG-TMV was heated for 18 h in ethylbenzene and subsequently characterized by AFM (e, f).

CHCl3 was removed by evaporation under reduced pressure. Subsequently, the initiator AIBN (2,20 -azobis(isobutyronitrile)) was added, and the reaction was heated at 60 C for 10 h (Figure 3). The solution was polymerized in a glass vial in order to cast the material. After reacting, a solid orange plastic cylinder was isolated by breaking the vial. Upon initial inspection, the radical polymerization did not appear to affect the linkage between the polymer and the protein, as evidenced by the persistent orange color of the azo linkage in the resulting material. This was confirmed by redissolving a small portion of the waxy polymer in chlorofom and recharacterizing the species after washing away the newly formed polymer. Figure 4c shows a TEM image stained with uranyl acetate of the redissolved material, which confirms that the plastic contained intact rods. We further demonstrated the thermal tolerance of the PEG-TMV by casting the plastic on the surface of a Roosevelt dime at 110 C. After reaching the glass transition temperature of polystyrene (∼95 C), the material re-formed to replicate the features of the dime. Again, after casting, the material retained the characteristic orange color of the azo bond that links the polymer chains to the capsid proteins. Finally, to ensure that the PEG-TMV was tolerant of the Langmuir 2010, 26(22), 17383–17388

Chemicals. Water (dd-H2O) used in biological procedures or as reaction solvents was deionized using a NANOpure purification system. Tobacco mosaic virus (TMV) was isolated after propagation in tobacco plants using a modified literature procedure,11 as described in the Supporting Information. This protocol routinely provided 1 g of purified virus for each kilogram of infected plant material. O-Methylpoly(ethylene glycol) was purchased from Nektar Theraputics. All other synthetic chemicals and solvents were of reagent grade and purchased from EMD and Aldrich Chemical Co. Instrumentation. DLS measurements were obtained using a Malvern Instruments Zetasizer Nano ZS and are reported as the “intensity” data for an average of three measurements. TEM images were obtained at the UC-Berkeley Electron Microscope Lab using a Philips FEI Tecnai 12 at 100 kV accelerating voltage. Tapping-mode AFM images were obtained in air on a Veeco multimode scanning probe microscope; samples were prepared on freshly cleaved muscovite mica. Sectional measurements were taken from height images using WSxM software.37 FPLC size exclusion chromatography was performed on a Bio-Rad BioLogic Duo-Flow using a GE Healthcare Sephacryl S-200 column. Synthesis of Virus-Polymer Conjugate Materials. The polymeric coating for TMV was synthesized in two steps from commercially available O-methylpoly(ethylene glycol) (MePEG), 2000 average molecular weight (Scheme 1).13 To ensure complete conversion, the primary alcohol of the Me-PEG was reacted twice under Mitsunobu conditions with N-hydroxyphthalimide, PPh3, and DIAD in CH2Cl2. The resulting alkoxyphthalimide was cleaved to the O-aminooxypoly(ethylene glycol) (PEG-ONH2) via hydrazinolysis. Preparation of the TMV capsid for attachment of the PEGONH2 was accomplished by installing >2000 copies of a ketone along the exterior of the virus rod. Native tyrosines at position 139 on the exterior of intact, wild-type TMV were modified under diazonium coupling conditions consisting of p-aminoacetophenone, NaNO2, and tosylic acid in water. An orange color developed during the course of this reaction due to the formation of the azo chromophore. The resulting viral rod (referred to as “k139-TMV”) has been extensively characterized previously.11 Synthesis of the virus-polymer conjugate material (PEG-TMV) (37) Horcas, I.; Fernandez, R.; Gomez-Rodrı´ guez, J. M.; Colchero, J.; GomezHerrero, J.; Baro, A. M. Rev. Sci. Instrum. 2007, 78, 013705.

DOI: 10.1021/la1039305

17387

Article was completed through the bioconjugation of PEG-ONH2 onto k139-TMV via oxime formation at pH 6.0 in aqueous buffer (Scheme 1). The extent of polymer incorporation was characterized by gel electrophoresis (Supporting Information Figure S2). Excess poly(ethylene glycol) was removed via size exclusion chromatography, and the solution was concentrated via centrifugal ultrafiltration (100 kDa MWCO). Transfer of Intact PEG-TMV to Organic Solvents. A twostep process was developed to transfer the PEG-TMV material into hydrophobic solvents. Representative images of a preparation on 0.1 g scale are shown in Supporting Information Figure S3. First, the PEG-TMV was precipitated from aqueous solution with 1.9 M (NH4)2SO4 and centrifuged in order to pellet the protein material. The supernatant was removed by decanting, and the bright orange pellet was redissolved in chloroform. The resulting solution was dried over sodium sulfate to remove any residual water and then filtered through glass wool. Second, the chloroform solution was diluted with a higher boiling solvent, and the chloroform was removed under reduced pressure. The dehydrated PEG-TMV material was then characterized by AFM on muscovite mica (Figure 1). Characterization of PEG-TMV in Organic Solvents. The size of PEG-TMV in chloroform, benzyl alcohol, diglyme (2ethoxyethyl ether), and o-dichlorobenzene was characterized using DLS, with the intensity data presented in Figure 2. Samples of PEG-TMV that were not heated were clarified by passage through a 0.45 μm PTFE syringe filter into a freshly cleaned and dust-free cuvette just before measuring. Samples were allowed to equilibrate in the instrument for 5 min at 25 C before collecting data. The integrity of PEG-TMV was also imaged by TEM. Samples were prepared by depositing organic solutions onto copper grids with an overlay of holey carbon. The material was then suspended over vacuum in the TEM for imaging.

17388 DOI: 10.1021/la1039305

Holder et al.

Thermal Stability of PEG-TMV. Solutions of PEG-TMV were tested for thermal stability by immersing a vial containing the material in an oil bath at either 120 or 160 C for 20 min. After cooling to room temperature, the samples were again characterized by DLS (without further 0.45 μm filtration to preserve any aggregates that might have formed) and TEM. Incorporation of PEG-TMV into Polystyrene. Chloroform containing PEG-TMV was diluted with an equal volume of neat styrene (0.5 mL), and the chloroform was removed under reduced pressure. Subsequently, AIBN was added to 7% w/v. The solution was then placed in a vial and heated at 60 C for 12 h to polymerize the styrene. Samples of solid polystyrene-TMV (PS-TMV) were isolated after cooling to room temperature. Thermal Casting of PS-TMV. The tolerance and malleability of the PS-TMV were tested by thermally casting the solid material (Figure 4). Slices of the PS-TMV cylinder were arranged on the substrate (the Roosevelt dime) and sandwiched between two stainless steel heat blocks at 110 C for 10 min. The material was allowed to cool to room temperature, after which it was imaged by digital photography through an optical microscope. Acknowledgment. This work was generously funded by the NSF (CHE-0449772). P.G.H. was supported by the UC Berkeley Chemical Biology Graduate Program (NRSA Training Grant 1 T32 GMO66698). The authors thank Dr. Harvey R. Johnson for helpful discussions. Prof. A. Paul Alivasatos, Prof. Peidong Yang, and Prof. Jean M. J. Frechet are acknowledged for the use of materials and instrumentation. They also thank the UC Berkeley Electron Microscopy Facility for guidance. Supporting Information Available: Full experimental details for all procedures, syntheses, and analyses. This material is available free of charge via the Internet at http://pubs.acs.org.

Langmuir 2010, 26(22), 17383–17388