Article pubs.acs.org/Langmuir
Pathways for Gold Nucleation and Growth over Protein Cages Ziyou Zhou,†,‡ Gregory J. Bedwell,§ Rui Li,§ Soubantika Palchoudhury,∥ Peter E. Prevelige,*,§ and Arunava Gupta*,†,‡ †
Center for Materials for Information Technology and ‡Department of Chemical and Biological Engineering, The University of Alabama, Tuscaloosa 35487, United States § Department of Microbiology, University of Alabama at Birmingham, Birmingham, Alabama 35294, United States ∥ Department of Civil and Chemical Engineering, University of Tennessee at Chattanooga, Chattanooga, Tennessee 37403, United States S Supporting Information *
ABSTRACT: Proteins are widely utilized as templates in biomimetic synthesis of gold nanocrystals. However, the role of proteins in mediating the pathways for gold nucleation and growth is not well understood, in part because of the lack of spatial resolution in probing the complicated biomimetic mineralization process. Self-assembled protein cages, with larger size and symmetry, can facilitate in the visualization of both biological and inorganic components. We have utilized bacteriophage P22 protein cages of ∼60 nm diameter for investigating the nucleation and growth of gold nanocrystals. By adding a gold precursor into the solution with preexisting protein cages and a reducing agent, gold nuclei/ prenucleation clusters form in solution, which then locate and attach to specific binding sites on protein cages and further grow to form gold nanocrystals. By contrast, addition of the reducing agent into the solution with incubated gold precursor and protein cages leads to the formation of gold nuclei/prenucleation clusters both in solution and on the surface of protein cages that then grow into gold nanocrystals. Because of the presence of cysteine (Cys) with strong gold-binding affinity, gold nanocrystals tend to bind at specific sites of Cys, irrespective of the binding sites of gold ions. Analyzing the results obtained using these alternate routes provide important insights into the pathways of protein-mediated biomimetic nucleation of gold that challenge the importance of incubation, which is widely utilized in the biotemplated synthesis of inorganic nanocrystals.
■
INTRODUCTION Studies related to nucleation and growth of gold nanocrystals (∼1−3 nm) are of significant interest for developing nextgeneration catalytic, sensor and nanoelectronic materials because they exhibit remarkable optical and electronic properties.1−4 Stabilization of such tiny nanocrystals against aggregation is the key for exploiting their unique properties, but the task has proved challenging. Ligands are generally used as stabilizers for the synthesis of gold nanocrystals in aqueous solution. Among these, the biological entities such as amino acids, DNA, and proteins can effectively regulate and direct the nucleation and growth of gold.5−8 Reported examples of using amino acids include negatively charged glutamic acid (Glu)9 and aspartic acid (Asp),7,10 positively charged arginine (Arg)11 and lysine (Lys),12 sulfur-containing cysteine (Cys),13 and amino acid with an imidazole functional group histidine (His)14 and tyrosine (Tyr).15 Biotemplates such as protein cages are an elegant assembly of multiple amino acid ligands packed within supramolecular structures with excellent monodispersity. The use of a biotemplate offers controlled growth of gold nanocrystals with a wide range of complex shapes and structures; however, a precise pathway for gold nucleation over them has thus far been elusive. This lack of understanding © 2017 American Chemical Society
is partly because of the limited spatial resolution in probing the complicated biomimetic mineralization process. Modifying the amino acid sequences in the biotemplates with Cys residues has been used as a key strategy to induce the binding of gold nanocrystals to the biotemplates.16 However, the formation of gold nanocrystals on unmodified protein cages has proved more challenging. To date, investigation on the pathways of gold nucleation over unmodified protein cages, visualizing both biological and inorganic components, has not been reported. Understanding the formation of gold nanocrystals on unmodified soft materials, such as protein cages, will be significant for developing new biohybrid materials with controlled plasmonic properties. We use the coat protein of bacteriophage P22 for templated nucleation and growth of gold nanocrystals. The rich chemical environment of P22 coat protein, which contains all 20 standard amino acid residues, offers abundant flexibility for binding, nucleation, and growth of gold nanocrystals.17−19 P22 coat proteins (420 copies) further self-assemble into a highly symmetric icosahedral cage with a dimension of around 60 Received: April 14, 2017 Published: May 18, 2017 5925
DOI: 10.1021/acs.langmuir.7b01298 Langmuir 2017, 33, 5925−5931
Article
Langmuir
UV−vis spectra were recorded before and after the addition of 5 μL of His or Cys solution (0.1 M) for 5 min. The hydrodynamic size and the surface charges of P22 protein cages and P22/Au in aqueous solution were evaluated using a Zetasizer Nano series dynamic light scattering (DLS) instrument. Simulation Method. The simulation was performed using the Python program (Python 2.7), taking different number of points with a size of 2.5 nm, distributed evenly on the surface of a sphere with a diameter of 60 nm. The number of points was determined using the number of Cys, His, Tyr, positively charged amino acid residues, and negatively charged amino acid residues.
nm.20 Different from amino acids or single protein molecules, the size and the symmetry of the assembled protein cage offer great convenience in the visualization of both biological and inorganic components. To investigate the P22 coat protein regulated nucleation of gold, we have adopted two different synthesis routes: (1) incubating P22 coat protein with gold precursors before the addition of the reducing agent, termed as R1, and (2) without incubation but adding gold precursors into protein solution with a preexisting reducing agent, termed as R2. The formation of gold nanocrystals was monitored by ultraviolet−visible (UV−vis) spectroscopy, coupled with transmission electron microscopy (TEM) observations. Our study provides valuable insights into possible pathways for gold mineralization over unmodified protein cages by visualizing both biological and inorganic components utilizing a combined UV−vis and TEM analysis protocol.
■
■
RESULTS AND DISCUSSION We have investigated the morphology and size of the assynthesized gold nanostructures, as well as P22 protein cages, using TEM. Figure 1 shows images of the assembled protein
EXPERIMENTAL SECTION
Materials. Gold was synthesized using commercially available reagents. Auric chloride (HAuCl4), sodium borohydride (NaBH4), Cys, and His were purchased from VWR (Pennsylvania, US). All chemicals were used as received, without any further purification. Protein Expression. A previously reported procedure was exactly followed for protein expression.21,22 Gold Synthesis. In the first route (R1) of biotemplated synthesis of gold, 0.1 mg of P22 coat protein cages was dispersed in 1 mL of deionized (DI) water at neutral pH in a plastic vial, into which 1.6 μL of 25 mM HAuCl4 was added, gently shaken for 1 min, and then allowed to incubate for a period of 1 h at room temperature before the addition of 8 μL of 25 mM NaBH4. In the second route (R2), 0.1 mg of P22 coat proteins was dispersed in 1 mL of DI water at neutral pH in a plastic vial, into which 8 μL of 25 mM NaBH4 was added, gently shaken for 1 min, and immediately followed by the addition of 1.6 μL of 25 mM HAuCl4. In both cases, the molar ratio of coat protein to gold precursor was about 20:1. Purification of P22/Au was achieved by centrifuging the solutions at 10 000 rpm for 10 min to remove the unbound gold nanocrystals in water. For gold nanocrystals synthesized without proteins, the amount and the ratio of the gold precursor and the reducing agent were kept the same. Materials Characterization. The morphology and the structure of the products were observed using TEM coupled with high resolution (Tecnai F-20). Uranyl acetate (2%) was used as a negative staining agent for imaging the assembled P22 coat protein cages. For in situ TEM measurements, 0.1 mg of P22 coat protein cages was incubated with 1.6 μL of 25 mM HAuCl4 in 1 mL of DI water at neutral pH in a plastic vial for 1 h. Subsequently, 5 μL of incubation solution was dropped onto a TEM grid, and the excess solution was blotted out using a filter paper after 2 min. The grid was washed three times with 10 μL of DI water to remove the unbound gold ions before allowing it to dry naturally. The electron-beam-induced in situ reduction and formation of gold nanocrystals on protein cages, which acted as gold ion reservoirs, were recorded using a Tecnai F-20 instrument with a low dose of 4 e−/Å2 at 80 keV to avoid sample damage. The optical measurements were recorded using a UV−vis spectrometer (Berkman DU 800). To determine the binding capability of P22 coat protein over gold ions, the absorption spectrum of 40 nmol/mL HAuCl 4 solution was measured using a UV−vis spectrometer before the addition of 0.1 mg (2.13 nmol) of P22 coat protein. After incubating the gold ions with protein for 1 h, a centrifuge filter device (MWCO 10 000) was used to collect the solution with unbound gold ions, which was again measured using the UV−vis spectrometer. The concentration of gold ions was determined using Beer−Lambert law, and the unbound gold ion concentration was estimated by subtracting the final molar amount from the initial molar amount of gold ions. For His and Cys displacement measurements, the purified gold nanocrystals synthesized with or without protein cages were dispersed in 1 mL of DI water inside of a quartz cuvette. The
Figure 1. TEM images of (a) P22 shells, (b) P22 shell-templated synthesis of gold with the incubation of gold precursors and biotemplates (route R1), (c) P22 shell-templated synthesis of gold with a preexisting reducing agent (route R2), (d) gold synthesized without the biotemplate. TEM samples were prepared after 5 min of reaction. Scale bars represent 100 nm in a−c. Scale bar represents 50 nm in d.
cages, gold nanocrystals over P22 templates, and gold nanocrystals synthesized without any templates. The assembly process and the resultant P22 coat protein cages are similar to those of wild-type P22 procapsid.23 Typically, 420 copies of the P22 coat protein (47 kDa) assemble with the aid of approximately 300 copies of scaffolding protein to form icosahedral T = 7 P22 cages (Figure 1a). On the basis of 50 randomly selected particles, the protein assembly is determined to have a diameter of 59.3 ± 3.2 nm (Figure S1). This is in good agreement with the reported size of wild-type P22 procapsid.21 Figure 1b,c presents the TEM images (without staining for better comparison) of Au/P22 nanostructures obtained using the synthesis routes R1 and R2, which have dimensions of 61.8 ± 4.8 and 62.3 ± 6.2 nm, respectively (the average size and the size distribution are detailed in Figure S1). Notably, the gold nanocrystals are positioned discretely over the protein cages, suggesting that there are selective sites for 5926
DOI: 10.1021/acs.langmuir.7b01298 Langmuir 2017, 33, 5925−5931
Article
Langmuir
Figure 2. UV−vis spectra of gold nanocrystal synthesis using different routes. Ratio of HAuCl4 to P22 coat protein (a) ∼20:1 and (b) ∼40:1. All measurements were taken after a reaction time of 1 min. The numbers indicate the wavelength, and the values in parenthesis indicate the absorbance.
the dependence of the gold precursor concentration by doubling the amount of gold precursors but keeping all other conditions unchanged, with the spectra shown in Figure 2b. Here, we observe plasmonic peaks in all three cases that are red-shifted, indicating the increase in size of the gold particles. Notably, the peak of gold produced in R2 is still the strongest in intensity and the widest, suggesting the most effective plasmonic interaction of the adjacent gold nanocrystals. Both TEM and UV−vis results demonstrate that controlled size and plasmonic properties of the gold nanocrystals can be obtained via the two different routes. To better understand the gold nucleation and growth, we have investigated how P22 coat protein cages interact with NaBH4 and HAuCl4, respectively. Figure S4 presents the UV−vis spectra of P22 cages before and after the addition of NaBH4. P22 coat proteins have a characteristic absorption peak at 280 nm, whereas NaBH4 has no absorption over the entire range of measurements (200− 800 nm). After the addition of NaBH4, the spectra in both cases are essentially the same. These results indicate that there is no significant interaction between P22 coat protein and the reducing agent NaBH4. The interaction of P22 coat protein and HAuCl4 has also been evaluated using UV−vis spectroscopy. HAuCl4 generates a characteristic absorption peak at 219 nm corresponding to Au3+, whereas P22 coat proteins strongly absorb light at 280 nm (Figure S5). When these two species are mixed together, the characteristic peaks of both P22 shells and HAuCl4 are no longer visible; instead, a broad peak emerges indicating a strong protein−HAuCl4 interaction. To roughly estimate the amount of gold precursor that binds with P22 shells, we have incubated an excess amount of HAuCl4 with P22 coat proteins (ratio of HAuCl4 to protein ≈ 20:1). We then evaluated the binding behavior by collecting the unbound gold precursors using a centrifuge filter device (MWCO 10 000), in which molecules larger than 10 000 are retained in the upper half, whereas lower molecular weight species can be recovered in the lower half of the device. The molecular weight of HAuCl4 is less than 400, whereas the molecular weight of a single P22 coat protein is 47,000. In principle, the P22 coat protein cages (molecular weight around 20 million) and/or P22−HAuCl4 complex will be retained in the upper half, whereas only unbound HAuCl4 can go down to the lower half of the tube. By employing this technique, we have been able to collect the solution with the unbound gold precursor and measure its concentration. As shown in Figure S5, the collected solution of the gold precursor shows the same absorption characteristics as that of the original
their growth and stabilization. The size of the individual gold nanocrystals has been measured using the ImageJ software, with the results shown in Figure S2. The size of the individual gold nanocrystals synthesized via R1 is 2.3 ± 0.6 nm, whereas that obtained via R2 is 3.1 ± 0.5 nm. In the absence of a biotemplate, the size of gold nanocrystals is 5.7 ± 1.5 nm. The gold particles formed on the protein cages show a statistically significant smaller size when compared with that without any template. This result indicates that P22 protein cages are effective templates for gold nanocrystal growth. Our TEM size analysis results agree with the DLS results, as shown in Figure S3. The hydrodynamic size of wild P22 protein cages is 69.7 nm. Au/P22 nanostructures obtained using the synthesis routes R1 and R2 have the hydrodynamic sizes of 72.4 and 73.1 nm, respectively. It is worth noting that the DLS results of Au/P22 nanostructures formed via R1 show a small peak at 2.9 nm, which may be attributed to the desorption of nonspecifically bound gold nanocrystals from the surface of the protein cages. We have monitored the formation of gold nanocrystals using UV−vis spectroscopy. In all cases, the volume of solution, the amount of gold precursors, the reducing agent, and the P22 coat protein cages are kept the same. However, the two routes result in remarkably different spectra for the same gold precursor to P22 coat protein ratio. As shown in Figure 2a, without any biotemplate, the reduction of HAuCl4 by NaBH4 (R1) results in an absorption peak at 522 nm, which is in agreement with the reported plasmonic peak of gold nanocrystals of similar size.24 For gold synthesis using R2, the nanocrystals exhibit an absorption peak at 516 nm. This blue shift of the plasmonic peak indicates a reduced size of the gold nanocrystals formed by biotemplating.25 Additionally, the absorption peak exhibited by biotemplated gold is broader than that of gold without any template, which commonly suggests a broader size distribution of nanocrystals.26,27 However, the TEM results have shown that the size distribution of biotemplated gold is actually narrower than that of gold without any template. Because the size distribution is not the cause for peak broadening, we reason that this phenomenon results from a strong plasmonic interaction of closely adjacent but discrete gold particles on the surface of P22 shells compared to that of gold particles freely distributed in aqueous solution.24,28 In the case of gold synthesized with gold precursor/template incubation, however, there is no notable plasmonic peak, which suggests that the size of gold under this synthesis condition is even smaller than that with R2, in agreement with TEM observations. We have also investigated 5927
DOI: 10.1021/acs.langmuir.7b01298 Langmuir 2017, 33, 5925−5931
Article
Langmuir
and stabilize the growth of gold, the potential sites can be as many as 17 640 over a single protein cage. In the simulation, however, 17 640 dots (2.5 nm) on a sphere (60 nm) form a very dense shell (Figure 3e) instead of discrete particles, which is completely different from the experimental results. In terms of negatively charged residues, with even more points (21 840) on a sphere, the simulated pattern will also not match with the TEM results. Earlier, based on the UV−vis measurements on the interaction of P22 coat protein and HAuCl4, we concluded that one protein cage is capable of binding 5040 Au ions, but even this value is high when compared with the simulated pattern (Figure 3d). On the basis of the close match between TEM and simulated pattern analyses, we conclude that the residues Cys or His likely regulate the growth of gold nanocrystals, although multiple residues can potentially bind with gold ions. Notably, Cys residues are not directly located on the outer surface of P22, suggesting that Au nanocrystals can form both on the outer surface and the inner surface of P22 coat protein cages. Previously, we have mentioned that the TEM image of protein−HAuCl4 is different from that of protein/gold in terms of spatial contrast. Although HAuCl4 can possibly positively stain the whole protein cage without any selectivity, the selectivity of the gold nanocrystals binds at some specific sites on the P22 coat proteins. To further confirm the selectivity of P22 coat protein for gold binding and growth, we have carried out in situ TEM observations. It is well-known that the noble metal ions such as Au3+ and Ag+ can readily be reduced using high energy radiation.31,32 Figure 4 shows the results of in situ TEM observations, in which the electron beam is utilized to trigger the reduction of Au ions and the formation of gold nanocrystals. It is worth noting that the spot size of the TEM beam for this experiment is decreased to minimize the sample
solution but with a remarkable decrease in the absorption peak. From the difference in peak intensity, we estimate that about 65% of gold precursor (around 26 nmol) binds with P22 coat proteins (2.13 nmol). We can, thus, conclude that there are roughly 12 gold ions per coat protein, corresponding to ∼5040 bound gold ions per cage. The interaction of a protein cage with HAuCl4 has also been confirmed using TEM measurements. As shown in Figure S6, spheres with size similar to P22 procapsid clearly indicate that heavy Au ions can bind and provide sufficient contrast for visualizing the protein cages, which are composed of light elements such as C, H, N, and S and are usually not clearly visible in regular TEM images without a staining agent. It should be noted that the TEM images of the protein cage/gold ions are different from that of the protein cage/gold nanocrystals in terms of the spatial contrast, indicating that the binding sites of gold nanocrystals are different from that of gold ions. A P22 coat protein has all 20 standard amino acid residues, including those that have previously been extensively utilized for the templated synthesis of gold nanocrystals: 1 copy of Cys, 2 copies of His, 8 copies of Tyr, 20 copies of Glu, 32 copies of Asp, 22 copies of Arg, and 20 copies of Lys. The number and the distribution of these residues have been analyzed using UCSF Chimera software,23,29 as shown in Figure S7. Do all of these residues act as nucleation sites for the binding, stabilizing, and growth of gold nanocrystals? To address this question, we have carried out simulations assuming that individual gold nanocrystals of size 2.5 nm grow uniformly over a spherical cage of 60 nm in diameter (based on gold formed via route R2).30 If Cys is the binding site for gold nanocrystals, there will be a maximum of 420 sites in a P22 shell for gold nucleation and growth. The simulated pattern is shown in Figure 3a, which
Figure 3. Size-dependent simulation of gold nanostructures. The dots are assumed to be 2.5 nm, and the spheres are assumed to be 60 nm. The number of dots on the sphere includes the following: (a) 420 dots, (b) 840 dots, (c) 3360 dots, (d) 5040 dots, (e) 17 640 dots, and (f) 21 840 dots.
is in good agreement with the TEM results of gold growth over P22 cages as shown in Figure 1c. If His acts as a binding site for gold, then 840 sites in a cage are potentially available for gold nucleation and growth. The simulated pattern is shown in Figure 3b, which is still in agreement with the TEM results of gold assembly over P22 cages. However, if Tyr residues provide nucleation sites for gold growth, there will be a maximum of 3360 gold nanocrystals grown over a protein cage. The simulated pattern in Figure 3c is certainly not well-matched with the TEM results. If positively charged residues mediate
Figure 4. In situ TEM observations of gold formation over P22 shells. (a) Before irradiation, (b) irradiation for 5 seconds, (c) irradiation for 10 seconds, and (d) irradiation for 15 seconds. Blue arrows indicate the particles shown in (c) that have disappeared in (d). Scale bar represents 20 nm in all images except in (a), where it represents 100 nm. 5928
DOI: 10.1021/acs.langmuir.7b01298 Langmuir 2017, 33, 5925−5931
Article
Langmuir
Figure 5. UV−vis measurements for the Cys-induced aggregation of gold without protein cages (a), synthesized with R1 (b), synthesized with R2 (c). As a comparison, the effect of His on gold aggregation is shown in (d).
tightly with proteins, whereas other particles with nonspecific binding are subject to desorption from the surface of protein cages and form aggregates induced by the presence of Cys. Surprisingly, for gold nanocrystals obtained via R2, the plasmonic peak just shifts slightly from 519 to 524 nm, showing a strong resistance to thiol displacement, indicating a strong bond between gold nanocrystals and coat proteins.31 The effect of His on gold aggregation is also tested. The UV− vis results of gold plasmonic peak, however, show no notable difference (Figure S7). Combining the results from both Cys and His displacement, we conclude that the binding sites for gold nucleation and growth in R2 are mainly provided by the Cys residues, whereas the multiple amino acid residues likely serve as binding sites for gold in R1. To further investigate the gold−protein interaction, we have carried out zeta potential measurements. As shown in Figure S8, the zeta potential (ζ) of P22 coat protein cages is −40.2 mV, which is in accordance with previous reported values.30 The high negative value of ζ potential indicates that the surface of P22 coat protein cages is dominated by negatively charged amino acid residues and good colloidal stability of P22 in aqueous solution. A colloidal system is generally considered well-stabilized via strong electrostatic interactions, if ζ < −30 mV or >30 mV.36 After the synthesis of gold nanocrystals in R1, the ζ potential value changes to −26.0 mV. The remarkable change of ζ potential suggests that the partial negatively charged amino acid residues have been covered by gold nanocrystals. However, full coverage of protein surface with gold nanocrystals does not occur. If the surface is fully covered by gold, then the ζ potential values should be in the range of −10 to −15 mV, which are the reported ζ potential values of pure gold nanocrystals synthesized without biotemplates.37 After the synthesis of gold in R2, the ζ potential slightly changes to −36.3 mV, suggesting that the most negatively
damage during TEM measurements. Figure 4a shows the morphology of P22 cage/HAuCl4. As expected, the protein cages are positively stained and can be visualized by the presence of Au ions, similar to the results shown in Figure S6. Five seconds after exposure, some particles emerge on the surface of P22 shells, suggesting the successful reduction of Au ions and the formation of Au particles. For longer electron exposure time (10 s), more particles are observed to randomly appear all over the P22 cages (Figure 4c). Interestingly, after exposure for 15 s, the number of particles decreases but some larger particles are formed (shown in Figure 4d) at the expense of smaller particles, which are indicated using blue arrows. This in situ TEM result suggests that the electron-beam-induced formation of gold clusters is nonpreferential. The P22 coat protein cages act as reservoirs that provide gold ions for the growth of gold nuclei or prenucleation clusters, starting at the binding sites of the gold ions on the proteins. However, the gold ion binding sites do not have a strong affinity for the gold nanocrystals, which tend to merge together and bind onto other specific sites on the protein cage. While preparing the revised article, Ueno and co-workers reported results on the movement of small gold nanocrystals to form large ones over crystalline protein cages,33 which is in accordance with our TEM observations (Figure 4c,d). We have further probed for the binding sites of gold on P22 coat protein cages using a Cys displacement method.34 Without protein templates, the addition of Cys can induce immediate aggregation of gold nanocrystals as evidenced by the remarkable shift of the plasmonic peak from 525 to 592 nm (Figure 5). This result is in agreement with previous reports on thiol-induced gold aggregation.35 For gold nanocrystals synthesized via R1, the plasmonic peak shows a moderate shift from 516 to 546 nm, on addition of Cys. This result suggests that a fraction of the gold nanocrystals binds very 5929
DOI: 10.1021/acs.langmuir.7b01298 Langmuir 2017, 33, 5925−5931
Article
Langmuir Notes
charged amino acid residues do not interact with the gold nanocrystals. On the basis of the above results and discussions, we are able to propose a formation mechanism of gold nanocrystals over P22 protein cages. In route R2 of gold synthesis, where the gold precursor is added into P22 solution with excess of preexisting reducing agent, the reaction immediately results in the formation of gold nuclei or prenucleation clusters, which find selective binding sites on the protein cages. The gold nuclei further grow to form larger nanocrystals, likely by particle− particle interactions, as shown from the larger particles and the few small particles after the exposure of 15 s (Figure 4). By contrast, for the R1 gold synthesis route, where proteins are bound with gold precursors, the gold nuclei are formed both at the binding sites and in the solution. However, gold nanocrystals formed on ion binding sites are not stable and tend to merge with particles formed on selective binding sites, based on the Cys displacement results.
The authors declare no competing financial interest.
■
REFERENCES
(1) Cobley, C. M.; Chen, J.; Cho, E. C.; Wang, L. V.; Xia, Y. Gold nanostructures: a class of multifunctional materials for biomedical applications. Chem. Soc. Rev. 2011, 40, 44−56. (2) Xie, J.; Zheng, Y.; Ying, J. Y. Protein-directed synthesis of highly fluorescent gold nanoclusters. J. Am. Chem. Soc. 2009, 131, 888−889. (3) Xu, C.; Wang, B.; Sun, S. Dumbbell-like Au−Fe3O4 nanoparticles for target-specific platin delivery. J. Am. Chem. Soc. 2009, 131, 4216− 4217. (4) Zhou, S.; McIlwrath, K.; Jackson, G.; Eichhorn, B. Enhanced CO tolerance for hydrogen activation in Au−Pt dendritic heteroaggregate nanostructures. J. Am. Chem. Soc. 2006, 128, 1780−1781. (5) Chaudhari, K.; Xavier, P. L.; Pradeep, T. Understanding the evolution of luminescent gold quantum clusters in protein templates. ACS Nano 2011, 5, 8816−8827. (6) Liu, C.-L.; Wu, H.-T.; Hsiao, Y.-H.; Lai, C.-W.; Shih, C.-W.; Peng, Y.-K.; Tang, K.-C.; Chang, H.-W.; Chien, Y.-C.; Hsiao, J.-K.; Cheng, J.-T.; Chou, P.-T. Insulin-directed synthesis of fluorescent gold nanoclusters: preservation of insulin bioactivity and versatility in cell imaging. Angew. Chem., Int. Ed. 2011, 50, 7056−7060. (7) Shao, Y.; Jin, Y.; Dong, S. Synthesis of gold nanoplates by aspartate reduction of gold chloride. Chem. Commun. 2004, 1104− 1105. (8) Das, S. K.; Dickinson, C.; Lafir, F.; Brougham, D. F.; Marsili, E. Synthesis, characterization and catalytic activity of gold nanoparticles biosynthesized with Rhizopus oryzae protein extract. Green Chem. 2012, 14, 1322−1334. (9) Polavarapu, L.; Xu, Q.-H. A single-step synthesis of gold nanochains using an amino acid as a capping agent and characterization of their optical properties. Nanotechnology 2008, 19, 075601. (10) Tan, Y. N.; Lee, J. Y.; Wang, D. I. C. Aspartic acid synthesis of crystalline gold nanoplates, nanoribbons, and nanowires in aqueous solutions. J. Phys. Chem. C 2008, 112, 5463−5470. (11) Xie, J.; Lee, J. Y.; Wang, D. I. C. Synthesis of single-crystalline gold nanoplates in aqueous solutions through biomineralization by serum albumin protein. J. Phys. Chem. C 2007, 111, 10226−10232. (12) Slocik, J. M.; Stone, M. O.; Naik, R. R. Synthesis of gold nanoparticles using multifunctional peptides. Small 2005, 1, 1048− 1052. (13) Kou, X.; Zhang, S.; Yang, Z.; Tsung, C.-K.; Stucky, G. D.; Sun, L.; Wang, J.; Yan, C. Glutathione- and cysteine-induced transverse overgrowth on gold nanorods. J. Am. Chem. Soc. 2007, 129, 6402− 6404. (14) Slocik, J. M.; Naik, R. R.; Stone, M. O.; Wright, D. W. Viral templates for gold nanoparticle synthesis. J. Mater. Chem. 2005, 15, 749−753. (15) Bhargava, S. K.; Booth, J. M.; Agrawal, S.; Coloe, P.; Kar, G. Gold nanoparticle formation during bromoaurate reduction by amino acids. Langmuir 2005, 21, 5949−5956. (16) Zahr, O. K.; Blum, A. S. Solution phase gold nanorings on a viral protein template. Nano Lett. 2012, 12, 629−633. (17) Qazi, S.; Liepold, L. O.; Abedin, M. J.; Johnson, B.; Prevelige, P.; Frank, J. A.; Douglas, T. P22 viral capsids as nanocomposite highrelaxivity MRI contrast agents. Mol. Pharm. 2013, 10, 11−17. (18) Patterson, D. P.; Prevelige, P. E.; Douglas, T. Nanoreactors by programmed enzyme encapsulation inside the capsid of the bacteriophage P22. ACS Nano 2012, 6, 5000−5009.
CONCLUSIONS We have investigated the pathways for gold nucleation and growth over P22 coat protein cages. Without incubation of gold precursors and biotemplates, the gold nanocrystals in the presence of preexisting reducing agent are formed via the direct attachment of gold nuclei and prenucleation clusters at the specific binding sites of proteins. With incubation of gold precursors and biotemplates, however, gold nucleation and growth occur in a concerted process: the addition of gold ions and the direct particle−particle interaction. Although the exact mechanism of how the gold nuclei find their specific binding sites remains elusive, the dynamic process of regulated nucleation of gold over P22 coat protein cages suggests the occurrence of a complex series of events that challenge the traditionally proposed mechanisms involving merely ion accretion. In biotemplated synthesis of inorganic nanocrystals, it is a common practice to incubate the precursors with biotemplates. However, this study shows that the incubation of gold precursors does not produce better outcome when compared with direct reduction without protein/gold precursor interaction. These insights will provide valuable guidance for the rational design of biotemplated fabrication of gold and the other inorganic nanostructures. ASSOCIATED CONTENT
* Supporting Information S
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.7b01298. Size distribution of P22-based particles as determined from TEM results; size distribution and average size of gold particles as determined from TEM results; DLS of P22 coat protein cages with and without gold; UV−vis spectra of P22 shells with and without NaBH4; comparison of TEM images of P22 coat protein cages; analyses of potential binding sites for gold on P22 coat protein using Chimera; and zeta potential data of P22 coat protein cages with and without gold (PDF)
■
ACKNOWLEDGMENTS
We thank Dr. Dongmao Zhang for helpful discussions and Dr. Yuping Bao for allowing us to use the zeta potential and DLS instruments. This work was supported by the US DOE, Office of Basic Energy Sciences, Div. of Material Sciences and Engineering Award DE-FG02-08ER46537.
■
■
■
AUTHOR INFORMATION
Corresponding Authors
*E-mail:
[email protected] (P.E.P.). *E-mail:
[email protected] (A.G.). 5930
DOI: 10.1021/acs.langmuir.7b01298 Langmuir 2017, 33, 5925−5931
Article
Langmuir (19) O’Neil, A.; Reichhardt, C.; Johnson, B.; Prevelige, P. E.; Douglas, T. Genetically programmed in vivo packaging of protein cargo and its controlled release from bacteriophage P22. Angew. Chem., Int. Ed. 2011, 50, 7425−7428. (20) Shen, L.; Bao, N.; Prevelige, P. E.; Gupta, A. Fabrication of ordered nanostructures of sulfide nanocrystal assemblies over selfassembled genetically engineered P22 coat protein. J. Am. Chem. Soc. 2010, 132, 17354−17357. (21) Bedwell, G. J.; Zhou, Z.; Uchida, M.; Douglas, T.; Gupta, A.; Prevelige, P. E. Selective Biotemplated Synthesis of TiO2 Inside a Protein Cage. Biomacromolecules 2015, 16, 214−218. (22) Zhou, Z.; Bedwell, G. J.; Li, R.; Bao, N.; Prevelige, P. E.; Gupta, A. P22 Virus-like Particles Constructed Au/CdS Plasmonic Photocatalytic Nanostructures for Enhanced Photoactivity. Chem. Commun. 2015, 51, 1062−1065. (23) Parent, K. N.; Khayat, R.; Tu, L. H.; Suhanovsky, M. M.; Cortines, J. R.; Teschke, C. M.; Johnson, J. E.; Baker, T. S. P22 coat protein structures reveal a novel mechanism for capsid maturation: stability without auxiliary proteins or chemical crosslinks. Structure 2010, 18, 390−401. (24) Zhang, T.; Wang, W.; Zhang, D.; Zhang, X.; Ma, Y.; Zhou, Y.; Qi, L. Biotemplated synthesis of gold nanoparticle−bacteria cellulose nanofiber nanocomposites and their application in biosensing. Adv. Funct. Mater. 2010, 20, 1152−1160. (25) Bastús, N. G.; Comenge, J.; Puntes, V. Kinetically controlled seeded growth synthesis of citrate-stabilized gold nanoparticles of up to 200 nm: Size focusing versus ostwald ripening. Langmuir 2011, 27, 11098−11105. (26) Tejamaya, M.; Römer, I.; Merrifield, R. C.; Lead, J. R. Stability of citrate, PVP, and PEG coated silver nanoparticles in ecotoxicology media. Environ. Sci. Technol. 2012, 46, 7011−7017. (27) Liu, Y.; Tsunoyama, H.; Akita, T.; Xie, S.; Tsukuda, T. Aerobic oxidation of cyclohexane catalyzed by size-controlled au clusters on hydroxyapatite: Size effect in the sub-2 nm regime. ACS Catal. 2011, 1, 2−6. (28) Su, K.-H.; Wei, Q.-H.; Zhang, X.; Mock, J. J.; Smith, D. R.; Schultz, S. Interparticle coupling effects on plasmon resonances of nanogold particles. Nano Lett. 2003, 3, 1087−1090. (29) Pettersen, E. F.; Goddard, T. D.; Huang, C. C.; Couch, G. S.; Greenblatt, D. M.; Meng, E. C.; Ferrin, T. E. UCSF ChimeraA visualization system for exploratory research and analysis. J. Comput. Chem. 2004, 25, 1605−1612. (30) Zhou, Z.; Bedwell, G. J.; Li, R.; Prevelige, P. E.; Gupta, A. Formation mechanism of chalcogenide nanocrystals confined inside genetically engineered virus-like particles. Sci. Rep. 2014, 4, 3832. (31) Schardein, G.; Donev, E. U.; Hastings, J. T. Electron-beaminduced deposition of gold from aqueous solutions. Nanotechnology 2011, 22, 015301. (32) Jiu, J.; Nogi, M.; Sugahara, T.; Tokuno, T.; Araki, T.; Komoda, N.; Suganuma, K.; Uchida, H.; Shinozaki, K. Strongly adhesive and flexible transparent silver nanowire conductive films fabricated with a high-intensity pulsed light technique. J. Mater. Chem. 2012, 22, 23561−23567. (33) Maity, B.; Abe, S.; Ueno, T. Observation of gold sub-nanocluster nucleation within a crystalline protein cage. Nat. Commun. 2017, 8, 14820. (34) Siriwardana, K.; Wang, A.; Vangala, K.; Fitzkee, N.; Zhang, D. Probing the effects of cysteine residues on protein adsorption onto gold nanoparticles using wild-type and mutated GB3 proteins. Langmuir 2013, 29, 10990−10996. (35) Vangala, K.; Ameer, F.; Salomon, G.; Le, V.; Lewis, E.; Yu, L.; Liu, D.; Zhang, D. Studying protein and gold nanoparticle interaction using organothiols as molecular probes. J. Phys. Chem. C 2012, 116, 3645−3652. (36) Xu, Y.; Qin, Y.; Palchoudhury, S.; Bao, Y. Water-Soluble Iron Oxide Nanoparticles with High Stability and Selective Surface Functionality. Langmuir 2011, 27, 8990−8997.
(37) Mu, H.; Liu, Q.; Niu, H.; Sun, Y.; Duan, J. Gold nanoparticles make chitosan−streptomycin conjugates effective towards Gramnegative bacterial biofilm. RSC Adv. 2016, 6, 8714−8721.
5931
DOI: 10.1021/acs.langmuir.7b01298 Langmuir 2017, 33, 5925−5931
