Selective Biotemplated Synthesis of TiO2 Inside a Protein Cage

Dec 15, 2014 - Department of Microbiology, University of Alabama at Birmingham, Birmingham, Alabama 35294, United States. ‡ Center for Materials and...
0 downloads 9 Views 4MB Size
Article pubs.acs.org/Biomac

Selective Biotemplated Synthesis of TiO2 Inside a Protein Cage Gregory J. Bedwell,† Ziyou Zhou,‡ Masaki Uchida,§ Trevor Douglas,§ Arunava Gupta,‡ and Peter E. Prevelige*,† †

Department of Microbiology, University of Alabama at Birmingham, Birmingham, Alabama 35294, United States Center for Materials and Information Technology, University of Alabama, Tuscaloosa, Alabama 35487, United States § Department of Chemistry, Indiana University, Bloomington, Indiana 47405, United States ‡

S Supporting Information *

ABSTRACT: Biological organisms have evolved tremendous control over the synthesis of inorganic materials in aqueous solutions at standard conditions. Such control over material properties is difficult to achieve with current synthesis strategies. Biotemplated synthesis of materials has been demonstrated to be efficient at facilitating the formation of various inorganic species. In this study, we employ a protein cage-based system to synthesize photoactive TiO2 nanoparticles less than 10 nm in diameter. We also demonstrate phase control over the material, with the ability to synthesize both anatase and rutile TiO2 using distinct biomineralization peptides within the protein cage. Finally, using analytical ultracentrifugation, we are able to resolve distinct reaction products and approximate their loading. We find that two distinct species comprise the reaction products, likely representing procapsid-like particles with early, precursor metal oxide clusters, and shells nearly full with crystalline TiO2 nanoparticles, respectively.



INTRODUCTION The study and use of viruses and virus-like particles (VLPs) has become a multidisciplinary field. Many viruses are robust, structurally and biophysically well characterized, and are reasonably tolerant of sequence and chemical modifications. As a result, these systems are readily tunable for a wide variety of applications and have sparked the interest of researchers in the fields of physics, biology, chemistry, materials science, and medicine. Interest in the use of protein cage nanoarchitectures as platforms for the synthesis of novel nanoparticles and nanomaterials has grown tremendously, with different systems affording different advantages (reviewed in refs 1−3). The impetus to use biological structures as nanoplatforms stems primarily from the ability to easily create particles with uniform dimensions and controllable physical properties. Both helically symmetric viruses as well as icosahedral viruses have been used successfully as nanoparticle platforms/scaffolds.1−3 Like the use of protein cage nanoarchitectures, biomineralization as a route for materials synthesis has seen a surge in popularity driven primarily by phage display technology, which allows investigators to “pan” for peptide sequences that catalyze the nucleation and growth of inorganic materials.4−12 The use of bio-organic templates to catalyze the nucleation and growth of inorganic materials offers an easy and environmentally friendly alternative to current chemical methods. Protein cage nanoarchitectures and biomineralization are well-suited companions, as has been previously demonstrated.11−16 Through the strategic placement peptide sequences within the framework of a protein cage, a material can, in principle, be synthesized in a controlled manner on any of the three surfaces inherent to a cage-like structure: the interior, the interface, and © 2014 American Chemical Society

the exterior. This offers unprecedented potential for the synthesis of inorganic nanoparticles in aqueous environments, as well as for the synthesis of hybrid material nanoparticles. The physical properties of many materials are also highly sensitive to factors such as size, shape, and phase.17,18 The use of protein cage nanoarchitectures can allow for enhanced control over these factors, making the properties of the material of interest readily controllable. The bacteriophage P22 represents an ideal candidate for use as a protein cage nanoarchitecture. P22 procapsid-like particles (PLPs) are T = 7 icosahedral structures with external and internal diameters of approximately 60 and 46 nm.19,20 The protein shell is composed of 420 copies of a single 47 kDa protein. PLPs are functional and stable as protein cage nanoarchitectures with or without the presence of an additional internal scaffolding protein.21 When present, the scaffolding protein is typically found at levels of approximately 100−300 copies per particle.21 Protein dissection experiments have revealed that the C-terminal ∼150 amino acids of the scaffolding protein binding (residues 141−303) are sufficient for coat protein binding. As such, fusions of peptides or proteins to this C-terminal region of the scaffolding protein can be readily internalized during PLP assembly or re-entering the shell after assembly has been completed. A recombinant system has been established for the generation of P22 PLPs, and large quantities of PLPs are easily obtainable on a benchtop scale.22 Along with our collaborators, we have shown various forms of Received: September 28, 2014 Revised: December 8, 2014 Published: December 15, 2014 214

DOI: 10.1021/bm501443e Biomacromolecules 2015, 16, 214−218

Article

Biomacromolecules

several hours. Stuffed shells were obtained by pelleting and resuspension in water or phosphate buffer. Mineralization Reaction. The mineralization reaction was carried out by diluting stuffed shells in either water or phosphate buffer (50 mM sodium phosphate pH 7.0; 50 mM NaCl) to a final concentration of 20 μM (relative to the coat protein). TiBALDH was then spiked into the solution to a final concentration of 10 mM. The reaction was allowed to sit at room temperature in the dark for up to 96 h. At each respective time point, samples were taken for light scattering without any additional processing. For all other experiments, the reaction products were run over a sephacryl S-300 column in water or dialyzed against Buffer B to remove any residual TiBALDH present in solution. Dynamic Light Scattering. Dynamic light scattering measurements were made in a PDDLS/Coolbatch (Varian, Inc.) at 25 °C in water. The scattered radiation was detected 90° from the incident light and the wavelength of light was 800 nm. A series of correlation functions were collected for each sample and averaged to remove noise. The averaged correlation function was then analyzed in MATLAB with an emulation of CONTIN called RILT (http:// www.mathworks.com/matlabcentral/fileexchange/6523-rilt). Transmission Electron Microscopy. TEM images were obtained using a Tecnai F-20 microscope operated at an acceleration voltage of 200 kV. Ten μL of sample solution was mounted on copper-Formvar grids, and excess solution was removed by filter paper. When appropriate, 2% uranyl acetate was added for staining. Analytical Ultracentrifugation. Sedimentation velocity experiments on intact particles were carried out in a Beckman XL-A ultracentrifuge at 25 °C. Detection of in tact particles was done at 280 nm. The particles were spun at 8000 rpm in an An60-Ti rotor. All runs were carried out in cells containing two-channel sectored centerpieces and sapphire windows. The data were analyzed using the ls-g*(s) model implemented in the program SEDFIT (http://www. analyticalultracentrifugation.com/default.htm). At the concentrations used, any contribution from unmineralized particles should be negligable. TiO2 particle size was determined by performing standard sedimentation velocity experiments in 6 M GuHCl at 25 °C and 15000 rpm. At the concentrations used, the absorbance of the protein should be negligible compared to the signal from TiO2. Particle size of the clusters was calculated from the observed s-value using the methods previously reviewed by Cölfen.34 This method makes the assumption that the particles are spherical and that the density of the particles is the same as the density of the bulk material. Photoactivity. The photocatalytic activity of P22-TiO2 was evaluated by photoreduction of methylene blue (MB, Sigma Chemicals Ltd.). An optical glass cuvette containing 1 mL of Buffer B with 10 μM MB and P22-TiO2 or empty P22 (0.05 mg or 0.25 mg) was purged with Ar and maintained sealed under this inert atmosphere. Each cuvette was illuminated with a Xe arc lamp (175 W, Lamba-LS, Sutter Instruments Co.) equipped with an IR filter (10 cm water filter). Photoreduction was monitored by the bleaching of the absorbance at 660 nm, analyzed with UV−vis spectroscopy (UV− vis, model 8453, Agilent). Particles used in the photoactivity experiments were synthesized in phosphate buffer and dialyzed against Buffer B to remove residual TiBALDH.

P22 to be useful as a protein cage-like structure for a variety of different applications, ranging from biomedical to materials science.2,11,15,23−28 TiO2 has been a well studied material for use in photocatalytic applications.29 Synthesis of TiO2 has been achieved using a variety of methodologies; however the use of protein cage nanoarchitectures for the synthesis of TiO2 is still being developed. Here we further demonstrate the modularity of P22 for materials synthesis through the internal biomineralization of TiO2 inside of P22 PLPs in aqueous conditions at neutral pH and ambient temperature. The mineralization reaction is driven by two previously reported peptide sequences engineered onto residues 141−303 of a scaffolding protein (herein referred to as T1 and T2, respectively).12,30−32 The T1 peptide sequence used in our studies, CHKKPSKSC, was previously shown to specifically bind to TiO2 with high affinity and, more recently, to efficiently nucleate the formation of TiO 2 from the TiBALDH precursor.30−32 The T2 peptide sequence, RKLPDAPGMHTW, was previously reported to nucleate the formation of anatase TiO2 from a TiBALDH precursor.12 When fused to the scaffolding protein, both peptides are readily internalized in the preformed protein cage.



MATERIALS AND METHODS

Cloning. The T1 scaffold construct was constructed using standard PCR methodologies. The TiO2 binding sequence was cloned onto the 5′ end of the scaffold gene in sequential PCR reactions. The template DNA used in the initial PCR reaction was a pET-3A plasmid containing the WT 141−303 gene. The T2 construct was constructed in a similar manner.12 Expression and Purification of Proteins. Wild-type P22 procapsids were expressed in a pET-based overexpression system in BL21 cells.22 Briefly, BL21 cells containing the gp8−5 assembly plasmid were grown in LB at 37 °C. Upon reaching mid log phase, expression of protein was induced with the addition of 1 mM IPTG. Induction was allowed to proceed for 3 h before harvesting. The cell pellets were resuspended in Buffer B (50 mM Tris-Cl pH 7.6; 25 mM NaCl) and frozen. Cells were lysed by freeze−thaw and sonicated. Following sonication, cell debris was cleared. The soluble material was then centrifuged through a 20% sucrose cushion and resuspended. Procapsid-like particles were then spun through a 5−20% sucrose gradient. The PLPs were found to migrate near the center of the gradient. The particles were removed and the scaffolding protein was extracted with 4 × 10 mL incubations with Buffer B (50 mM Tris-Cl pH 7.6; 25 mM NaCl) supplemented with 0.5 M GuHCl. Shells were stored at 4 °C in Buffer B. The individual T1 and T2 constructs were expressed and lysed in 3 M GuHCl; 25 mM NaHCO3, pH 10. Following cell lysis, the cell debris was separated from the soluble material by high-speed spin. The soluble fraction is then concentrated and loaded onto a S-200 sephacryl column equilibrated in 1 M NaCl; 25 mM NaHCO3, pH 10, flowing at 0.5 mL/min. The peak fractions are pooled and diluted to a final NaCl concentration of 200 mM and run over a HiTrap Q sepharose column. The flow-through is collected and GuHCl is added to a final concentration of 3 M. The material is concentrated and filtered and applied again to an S-200 column. The peak fractions containing T1 are identified SDS-PAGE and the fractions are pooled, concentrated, and stored at −80 °C until use. Re-entry. Scaffolding protein has been previously shown to be able to re-enter P22 shells and stably bind the coat protein upon incubation of the scaffolding protein and the empty procapsid together.33 DTT was added to a 10 mM final concentration. Re-entry of the purified T1 scaffolding protein into empty shells was achieved either by incubation of the two components at a ≥ 2:1 T1/coat molar ratio. The components were allowed to incubate either at room temperature for



RESULTS While both the T1 scaffolding construct and TiBALDH are soluble on their own, we observed precipitation upon mixing the two. In an attempt to better control the reaction and to keep the mineralized material soluble in an aqueous environment, we confined the T1 scaffolding protein to the interior of the P22 shells.35 Prior to mineralization, the appearance of these particles under the microscope was typical of particles containing only coat and scaffold proteins (data not shown). To assess whether or not mineralization occurred, mineralized particles were imaged by transmission electron microscopy (TEM). In Figure 1A, 2% uranyl acetate stained images of the 215

DOI: 10.1021/bm501443e Biomacromolecules 2015, 16, 214−218

Article

Biomacromolecules

Given the strong dependence of scattering intensity on particle size, these data are good evidence that no external mineralization and/or particle aggregation is occurring during the course of the reaction. High-resolution TEM analysis of particles containing the T1 scaffolding construct from 24 h (Figure 2A) and 96 h (Figure

Figure 1. (A) Uranyl acetate (2%) stained T1 particles from the mineralization reaction after 24 h reaction. Particles with P22 morphology and size are readily discerned in the micrograph. The dark cores suggest the presence of electron dense material, possibly attributable to TiO2. (B) Unstained image of the same sample imaged in A showing a discrete cluster of electron dense material with an average radius very close to that of the P22 core, favoring the presence of TiO2 inside of the particle cores. The material is not monolithic, but rather appears to be a composite of smaller nanoparticles. (C) DLS particle size distributions for the T1 starting material (unmineralized shells) and shells incubated up to 96 h with TiBALDH. No change in the apparent particle size is observed, confirming that the mineralization reaction is occurring on the interior of the PLPs.

Figure 2. (A) HRTEM image of a portion of a mineralized particle with the T1 construct following a 24 h incubation with TiBALDH in water. The image shows the presence of crystalline material with a dspacing corresponding to the (110) plane of rutile TiO2. (B) HRTEM image of a portion of a particle with the T1 construct following a 96 h incubation with TiBALDH in water. Like in (A), the image shows the presence of crystalline material. This image, however, shows two distinct sets of lattice fringes, corresponding well to the (110) and (111) planes of rutile, TiO2. (C) A mineralized particle with the T1 construct synthesized after 24 h in phosphate buffer. The d-spacing of this material is identical to that observed after 24 h in water. (D) A particle synthesized in phosphate buffer with the T2 construct, previously shown to form anatase TiO2. Consistent with previous data, the d-spacing observed in this particle corresponds well to the (101) plane of anatase TiO2, confirming the role of the peptide sequence in crystal phase determination in our system. Whole-particle images from which (C) and (D) are derived are shown in Supporting Information, Figure 1A and B, respectively.

reaction products after 24 h show particles with a similar size and morphology to that of typical P22 procapsids, indicating that incubation of the particles with TiBALDH does not result in the disruption of the protein cage. The average outer and inner radii of these particles were determined to be approximately 32 and 24 nm, respectively, similar to the values determined by cryo-EM reconstructions of P22 PLPs.20 The core of the stained particles is dark, suggesting the presence of electron dense material inside of the shells, though unequivocally attributing this to the presence of TiO2 and not stain penetration is dubitable. In Figure 1B, an unstained image from the same sample as imaged in Figure 1A clearly shows electron dense material and analysis of several of these particles yields an average radius of 25 nm, a value very close to that of the inner radius of the particles seen in 1A. These data strongly support the presence of titanium localized to the interior of the P22 PLP. The distribution of the dense material in these particles is irregular, suggesting that the synthesized material is not monolithic, and likely arises from multiple nucleation events. We used dynamic light scattering (DLS) to assess the structural integrity and size homogeneity of the mineralized particles. (Figure 1C). The intensity weighted particle size distribution for the starting material (0 h) and over the course of 4 days (24−96 h) does not change over time, and the measured radius of 34 nm is in very good agreement with the radius measured from both stained micrographs and cryo-EM reconstuctions.

2B) samples show the presence of lattice fringes, indicating the presence of crystalline or partially crystalline material. The dspacing of the two different observed fringes was determined to be 3.19 and 2.16 Å, corresponding very well with the known dspacing of the (110) and (111) planes of rutile TiO2, respectively (JCPDS PDF# 88−1175). Similar d-spacings were observed when the reaction with the T1 particles was carried out in phosphate buffer (Figure 2C). To interrogate the role of the engineered peptide sequence in the phase determination of TiO2, we mineralized particles using the T2 peptide fusion protein. The T2 sequence was previously reported to nucleate formation of anatase TiO2 in phosphate buffer12 and the T2 fusion protein did so in the context of our system (Figure 2D), as evidenced by the observed d-spacings of 3.51 Å, which correspond very well to the (101) plane of anatase (JCPDS PDF# 86−1157). Whole-particle HRTEM images of the particles shown in Figure 2C,D are included in the Supporting Information as Figure S1A and B, respectively. 216

DOI: 10.1021/bm501443e Biomacromolecules 2015, 16, 214−218

Article

Biomacromolecules To test the photoactivity of the reaction products, we measured the photoreduction of methylene blue (MB) in solution with particles containing the T1 scaffolding construct after a 24 h mineralization reaction.36 Mineralized particles showed enhanced photoactivity relative to unmineralized particles, as shown in Figure 3A.

Figure 3. Analysis of the rate of methylene blue photoreduction in the presence of mineralized or unmineralized shells upon illumination with light. The data show an increase in the rate of light-induced methylene blue photoreduction in the presence of mineralized shells compared to unmineralized shells, demonstrating the photoactivity of the synthesized particles. The photoactivity is also a good indicator of the presence of crystalline material. Figure 4. (A) Combined sedimentation coefficient and particle density distributions after 0 to 96 h incubation with TiBALDH. Two dominant species are observed in all of the reaction products. (B) Particle size distribution of the TiO2 nanoparticle synthesized inside of P22 PLPs. The synthesized nanoparticles are small and do not increase significantly in size, even after 96 h.

The micrograph in Figure 1B suggests that the TiO2 particles in the P22-based system are small and not monolithic in nature. To analyze the size of the synthesized inorganic nanoparticles, mineralized T1-containing particles were disrupted in 6 M GuHCl to release the inorganic nanoparticles into solution. The measured sedimentation coefficient distribution of the inorganic particles was then converted into an apparent particle size distribution (Figure 4B). 34,37 The calculated size distribution shows the presence of several discrete species up to approximately 10 nm in diameter, confirming that the nanoparticles are not monolithic and that the particle sizes are small and reasonably well-defined. The observation of discrete nanoparticle species in the inorganic particle size distributions has been attributed to defined growth at particular faces of the crystal.37 We hypothesize that the discrete TiO2 species that we observe are the result of a similar physical phenomena. From the sedimentation coefficient distributions of nondisrupted T1-containing reacted shells in Figure 4A, it is apparent that the biomineralization process in our system is complex. The distributions are biphasic, with populations sedimenting at 221 and 363 S, respectively. The 221 S species is distinct from the unmineralized population, centered at 211 S. It is possible that the 221 S species represents a stable early intermediate responsible for the formation of rutile TiO2 at each nucleation site. This, in turn, would make the 221 S species a necessary precursor for the 363 S species. This is consistent with the fact that the fraction of particles in the 221 S species remains relatively constant over time, while the 363 S fraction increases. This can only be the case if the 221 S population is both being consumed and being made simultaneously. The 363 S species appears to represent the end-point of TiO2 incorporation into particles. The size of the synthesized inorganic particles does not change appreciably over time. However, the intensity of the 363 S species does

over the first 72 h. These observations suggest a mechanism in which the nucleated species at each scaffolding protein grow independently of other nanoparticles.



CONCLUSIONS In this study, we employ the P22-based biotemplating system for the synthesis of crystalline TiO2. We observe very good spatial control over the location of mineralization, as evidenced by the TEM and DLS data in Figure 1, further demonstrating that the P22-based system is well suited for the spatially constrained synthesis of inorganic materials. The crystalline material synthesized with the T1 constructed appears to be rutile TiO2. Previous work done with the T1 peptide in solution found that the peptide itself was capable of driving the formation of TiO2, but that the material was amorphous prior to calcination.32 The disparity between these results highlights apparent differences between biomineralization in free solution and within a constrained environment. The physical basis for the differences, however, is unknown. Our ability to synthesize anatase TiO2 with the T2 construct suggests that the observed phase discrimination in our system is in fact peptide-dependent. Moreover, it demonstrates a level of control over TiO2 synthesis previously unrealized in a biotemplated synthesis system. The photoactivity of the material synthesized with the T1 construct acts to both corroborate the presence of crystalline TiO2 and demonstrate potential practical applications for the system. The sedimentation experiments provide a glimpse into the mineralization process in our system. We 217

DOI: 10.1021/bm501443e Biomacromolecules 2015, 16, 214−218

Article

Biomacromolecules

W.; Baker, D.; King, J.; Chiu, W. Proc. Natl. Acad. Sci. U.S.A. 2010, 108, 1355. (21) Prevelige, P. E.; Thomas, D.; King, J. J. Mol. Biol. 1988, 202, 743. (22) Kang, S.; Prevelige, P. E., Jr. J. Mol. Biol. 2005, 347, 935. (23) Lucon, J.; Qazi, S.; Uchida, M.; Bedwell, G. J.; LaFrance, B.; Prevelige, P. E.; Douglas, T. Nat. Chem. 2012, 10, 781. (24) O’Neil, A.; Reichhardt, C.; Johnson, B.; Prevelige, P. E.; Douglas, T. Angew. Chem., Int. Ed. 2011, 50, 7425. (25) Uchida, M.; Morris, D. S.; Kang, S.; Jolley, C. C.; Lucon, J.; Liepold, L. O.; LaFrance, B.; Prevelige, P. E., Jr.; Douglas, T. Langmuir 2012, 28, 1998. (26) Patterson, D. P.; Prevelige, P.; Douglas, T. ACS Nano 2012, 6, 5000. (27) Qazi, S.; Liepold, L. O.; Abedin, M. J.; Johnson, B.; Prevelige, P.; Frank, J. A.; Douglas, T. Mol. Pharmaceutics 2012, 10, 11. (28) Zhou, Z.; Bedwell, G. J.; Li, R.; Prevelige, P. E.; Gupta, A. Sci. Rep. 2014, 4, 3832. (29) Fujishima, A.; Honda, K. Nature 1972, 228, 37. (30) Chen, H.; Su, X.; Neoh, K. G.; Choe, W. S. Anal. Chem. 2006, 78, 4872. (31) Chen, H.; Su, X.; Neoh, K. G.; Choe, W. S. Langmuir 2009, 25, 1588. (32) Choi, N.; Tan, L.; Jang, J. R.; Um, Y. M.; Yoo, P. J.; Choe, W. S. J. Inorg. Biochem. 2012, 115, 20. (33) Parker, M. H.; Casjens, S.; Prevelige, P. E. J. Mol. Biol. 1998, 281, 60. (34) Planken, K. L.; Colfen, H. Nanoscale 2010, 2, 1849. (35) Parker, M. H.; Brouillette, C. G.; Prevelige, P. E., Jr. Biochemsitry 2001, 40, 8962. (36) Mills, A.; Wang, J. J. Photochem. Photobiol. A 1999, 127, 123. (37) Cölfen, H.; Pauck, T. Colloid Polym. Sci. 1997, 275, 175.

observe two distinct populations with s-values different from the unreacted control. One of these populations appears to be made and consumed throughout the course of the reaction, while the other appears to represent the end point of the reaction. It is unclear exactly what metal-oxide species occupy the intermediate population we observe. However, it does appear that the mineralization reaction in our system is more complex than a simple nucleation-limited growth process.



ASSOCIATED CONTENT

S Supporting Information *

Whole particle images of the particles shown in Figure 2C,D. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This research was supported by the U.S. Department of Energy, Office of Basic Energy Sciences, Division of Materials Sciences and Engineering under Award # DE-FG02-08ER46537.



REFERENCES

(1) Douglas, T.; Young, M. Nature 1998, 393, 152. (2) Kang, S.; Uchida, M.; O’Neil, A.; Li, R.; Prevelige, P. E.; Douglas, T. Biomacromolecules 2010, 11, 2804. (3) Steinmetz, N. F. Mol. Pharmaceutics 2013, 10, 1. (4) Kharlampieva, E.; Slocik, J. M.; Singamaneni, S.; Poulsen, N.; Kröger, N.; Naik, R. R.; Tsukruk, V. V. Adv. Funct. Mater. 2009, 19, 2303. (5) Kharlampieva, E.; Tsukruk, T.; Slocik, J. M.; Ko, H.; Poulsen, N.; Naik, R. R.; Kröger, N.; Tsukruk, V. V. Adv. Mater. 2008, 20, 3274. (6) Kroger, N.; Dickerson, M. B.; Ahmad, G.; Cai, Y.; Haluska, M. S.; Sandhage, K. H.; Poulsen, N.; Sheppard, V. C. Angew. Chem., Int. Ed. 2006, 45, 7239. (7) Nam, Y. S.; Magyar, A. P.; Lee, D.; Kim, J. W.; Yun, D. S.; Park, H.; Pollom, T. S., Jr.; Weitz, D. A.; Belcher, A. M. Nat. Nanotechnol. 2010, 5, 340. (8) Nonoyama, T.; Kinoshita, T.; Higuchi, M.; Nagata, K.; Tanaka, M.; Sato, K.; Kato, K. J. Am. Chem. Soc. 2012, 134, 8841. (9) Zhao, C.-X.; Yu, L.; Middelberg, A. P. J. RSC Adv. 2012, 2, 1292. (10) Aizenberg, J. Adv. Mater. 2004, 16, 1295. (11) Shen, L.; Bao, N.; Prevelige, P. E.; Gupta, A. J. Am. Chem. Soc. 2010, 132, 17354. (12) Schoen, A. P.; Schoen, D. T.; Huggins, K. N.; Arunagirinathan, M. A.; Heilshorn, S. C. J. Am. Chem. Soc. 2011, 133, 18202. (13) Klem, M. T.; Young, M.; Douglas, T. J. Mater. Chem. 2008, 18, 3821. (14) Klem, M. T.; Young, M.; Douglas, T. J. Mater. Chem. 2010, 20, 65. (15) Reichhardt, C.; Uchida, M.; O’Neil, A.; Li, R.; Prevelige, P. E.; Douglas, T. Chem. Commun. (Cambridge, U.K.) 2011, 47, 6326. (16) Kang, S.; Jolley, C. C.; Liepold, L. O.; Young, M.; Douglas, T. Angew. Chem., Int. Ed. 2009, 48, 4772. (17) Burda, C.; Chen, X.; Narayanan, R.; El-Sayed, M. A. Chem. Rev. 2005, 105, 1025. (18) El-Sayed, M. A. Acc. Chem. Res. 2004, 37, 326. (19) Parent, K. N.; Khayat, R.; Tu, L. H.; Suhanovsky, M. M.; Cortines, J. R.; Teschke, C. M.; Johnson, J. E.; Baker, T. S. Structure 2010, 18, 390. (20) Chen, D. H.; Baker, M. L.; Hryc, C. F.; DiMaio, F.; Jakana, J.; Wu, W.; Dougherty, M.; Haase-Pettingell, C.; Schmid, M. F.; Jiang, 218

DOI: 10.1021/bm501443e Biomacromolecules 2015, 16, 214−218