Journal of Colloid and Interface Science 418 (2014) 31–36
Contents lists available at ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
Control of gold nanoparticles based on circular DNA strand displacement Cheng Zhang a, Jingjing Ma a,1, Jing Yang b, Yafei Dong c,⇑, Jin Xu a,⇑ a Institute of Software, School of Electronics Engineering and Computer Science, Key Laboratory of High Confidence Software Technologies of Ministry of Education, Peking University, No. 5 Yiheyuan Road, Haidian District, Beijing 100871, People’s Republic of China b School of Control and Computer Engineering, North China Electric Power University, No. 2 Beinong Road, Huilongguan, Changping District, Beijing 102206, People’s Republic of China c College of Life Science, Shannxi Normal University, 199 South Chang’an Road, Xi’an 710062, People’s Republic of China
a r t i c l e
i n f o
Article history: Received 6 September 2013 Accepted 28 November 2013 Available online 14 December 2013 Keywords: DNA strand displacement Circular DNA Gold nanoparticles DNA self-assembly TEM image
a b s t r a c t In this study, DNA strand displacement is utilized to control the aggregation of DNA/gold nanoparticles (AuNPs) based on circular DNA, in which DNA/AuNP conjugates are captured and released by adding different DNA signal strands. Using this strategy, single DNA/nanoparticle building blocks are capable of assembling into complex structures of two and three circular DNA/nanoparticles. The existence of these structures is demonstrated by gel electrophoresis and transmission electron microscopy (TEM) analysis. This advance has potential applications in controlling, transporting and detecting DNA/AuNP conjugates with subsequent manipulation of the structure and function of these assemblies. Ó 2013 Elsevier Inc. All rights reserved.
1. Introduction Programmable gold nanoparticle (AuNP) assembly leads to great advances in nano-construction. DNA-mediated self-assembly of AuNP is a promising way to organize and control a complicated nano-system. Through programmable DNA self-assembly, positions and geometries of AuNPs can be controlled precisely and hierarchically. A critical feature that contributes to the connection of DNA and AuNP is specific binding recognition (double-strand hybridization): that the attached DNA strands (DNA molecules) play a role as specific connection, which endow the homogenous AuNPs with different binding labels. A series of research works based on well-defined DNA/gold nanoparticle architectures have been established taking advantage of DNA base pairing [1–6]. Recently, flexible and deformable DNA nano-systems have been constructed, in which the method of strand displacement is used frequently, which is a mechanism of self-driving hybridization with DNA molecules branching migration [7–12]. By controlling several factors such as toehold sites (initial hybridizing region), strand lengths and hybridized structures, the DNA strands with the longer hybridizing regions can displace those with the shorter. In fact, the strand displacement method has been utilized in engineering many kinds of synthetic DNA machines [13–15]. It
⇑ Corresponding authors. E-mail addresses: [email protected] (C. Zhang), [email protected] (J. Ma), [email protected] (J. Yang), [email protected] (Y. Dong), DNA@ pku.edu.cn (J. Xu). 1 This author contributed equally with the first author. 0021-9797/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcis.2013.11.079
demonstrated that such biomolecular autocatalysis displacement can be driven by the free energy of specific base-pair (bp) hybridization, to make free energy of system to stable states. It has been a great challenge to fabricate reversible and complicated nanodevices based on gold nanoparticles. However, combined with DNA strand displacement, DNA/AuNP nanodevices capable of releasing and deforming functions have been developed [16–22]. In fact, many approaches for such applications have been proposed in recent years [21]. Weihong Tan’s group developed a Boolean logic operation in response to two pro-angiogenic targets in the area of cancer diagnosis and treatment [22]. In general, there are two main methods to prepare DNA/AuNP conjugates, both of which rely on thiolated DNA. One is functionalizing a single AuNPs with many DNA strands, which are randomly bound to the surface of particles [23]. Another is building AuNPs with a discrete number of single-stranded DNA [24]. Depending on different applications, numbers of oligonucleotides could be precisely connected on an individual nanoparticle [25]. Although prior work has advanced the field DNA/AuNP assembly, most examples that incorporated strand displacement capability were based on short linear DNA structures, thus limiting assemblies of more complicated hierarchical molecular scaffolds. In addition, many DNA/AuNP assemblies have been established choosing a strategy of settled structures, in which once the hybridizations complete, the positions and geometries of AuNPs will not change. Thus, the challenge of engineering nanodevices with tunable structures has attracted more interest [26–30]. In this study, we established a nanodevice based on DNA/AuNP conjugates displacement, in which DNA/AuNP conjugates were captured and
32
C. Zhang et al. / Journal of Colloid and Interface Science 418 (2014) 31–36
released selectively by adding different DNA signal molecules. The novel feature of our assemblies is the use of a circular DNA strand as a template on which to assemble DNA/AuNP conjugates, providing opportunities for construction of structurally more complex systems than has been possible before.
used filter paper to wick off excess buffer. After repeating washing for 1–2 times, grids were air-dried for analyzing (Supporting information Section 8).
2. Experimental materials and methods
As shown in Fig. 1, a nanodevice uses circular single-stranded DNA as a basic subunit, which consists of three regions, providing hybridization sites for DNA/AuNP conjugates NP1, NP2 and NP3. We also synthesized DNA/AuNP conjugates called AuDP2 and AuDP3, which are complementary to NP2 and NP3, and therefore are capable of displacing their respective partners from the circular DNA scaffold. After specific hybridization, three regions capture NP1, NP2 and NP3 by formation of double strand DNA lengths of 19 bp, 23 bp and 17 bp. Thus, three AuNPs are attached to one DNA circle to form a supramolecular assembly. In the design principle, the sequences of NP1, NP2 and NP3 all consist of three regions, including a flexible oligo-thymine domain to act as a spacer to extend the base pairing region away from the NP surface, a capturing domain complementary to a specific region of the circular template, and a 30 -terminal extension (Fig. S1). In addition, the DNA strands in NP2 and NP3 contained toehold domains between the spacer and recognition domains. These short sequences provide nucleation sites for the displacement strands used to disassemble the nanostructures. In the initial capture step, the DNA circle hybridizes with the complementary capture domain of the single DNA strands on DNA/AuNP conjugates NP1, NP2 and NP3. Then, to verify whether the DNA nanodevice can function as designed, a releasing course is implemented via strand displacement, in which toehold regions on NP conjugates will uniquely recognize single DNA strands on conjugates AuDP2 and AuDP3. Notably, a length of thymine bases is
2.1. Materials The main materials and chemicals used in our experiments are as follows: gold nanoparticles (AuNPs) with diameter of 15 nm from Ted Pella; Bis(p-sulfonatophenyl)phenylphosphinedipotassium salt dihydrate (BSPP) from Strem Chemical; Tris–Borate-EDTA buffer (TBE) and Tris–Acetic-EDTA buffer (TAE) from Solarbio; disulfide protected thiolated oligonucleotides by purification of HPLC from Sangon China; agarose from BIOWEST; methylenebisacrylamide and acrylamide monomer from TCI, Stains-All from SIGMA ALORICH. 2.1.1. Circularization of strand C DNA annealing (hybridization) and T4 ligase were used to form the DNA circle. The reactants strands C and CP (see Supporting information, Table S1 for sequences) were mixed together in TBE buffer and provided an annealing temperature conditions which were: 95 °C for 4 min, 65 °C for 30 min, 50 °C for 30 min, 37 °C for 30 min, 22 °C for 30 min. Then the solution was incubated with T4 ligase at 16 °C for 10 h. After ligation strand CP1 was added to hybridize with strand CP leaving the DNA circle free. 2.1.2. PAGE gel purification of DNA circle To separate DNA circle from other products non-denaturing polyacrylamide gel was used. The specific product of DNA circles were purified by directly cutting the target band and obtained by overnight alcohol precipitation at 20 °C (Suppporting information Section 3).
3. Results and discussion
2.1.3. Phosphination of AuNPs Solid powders of Bis(p-sulfonatophenyl) phenylphosphinedihydratedipotassium salt (BSPP) and AuNPs solution were mixed together and either shaken or stirred overnight at room temperature. Then NaCl (solid) was added slowly into the mixture. When the color of the solution changed from red to purple, it was centrifuged for 15 min and the liquid was removed. Finally, the AuNPs were dissolved in 1 BSPP solution. 2.1.4. Preparation of AuNPs-DNA conjugates DNA strands modified by disulfide were incubated with the AuNPs produced above for ligation (Suppporting information Section 6). Then 4% agarose gel was used to separate the AuNPsDNA conjugates. The AuNPs-DNA conjugates (NP and AuDP) were collected by using a centrifugal filter device. 2.1.5. Hybridization and displacement of AuNPs-DNA conjugates The collected NP conjugates were hybridized with DNA circles for 12 h at room temperature. Then AuDP conjugates were added into the solution for displacement of NP for 12 h at room temperature. The products were then electrophoresed in 4% agarose gels. 2.1.6. TEM images TEM images were obtained using a Tecnai G20. To increase surface hydrophilicity, TEM grids were ionized for 20–30 s using oxygen plasma cleaner (Harrick Plasma PDC-32). Then 2–4 lL of purified sample was spotted on the surface and left on the grid for 3 min. Later, added 10 lL of 0.5 TBE buffer for washing and
Fig. 1. Scheme of the assembly and release process. (a) Hybridizing to assemble into the three-particle structure. NP1, 2 and 3 indicate DNA/AuNP conjugates. (b) Displacement with DNA/AuNP conjugates. AuDP2 and 3 hybridize with NP2 and 3, leaving NP1 on the DNA circle. (c) Displacement with DNA strands, DNA/AuNP conjugates NP2 and 3 are removed from the DNA circle.
C. Zhang et al. / Journal of Colloid and Interface Science 418 (2014) 31–36
utilized to add sufficient flexibility and reduce straining of the structures, which would facilitate a flexible DNA strand structure turning during the displacement reaction. Similarly, at the far end of a single strand on the AuNP, another DNA region, also consisting of thymine bases, is designed for sequence length extension. In the same way, conjugates AuDP2 and AuDP3 also consist of hybridization, toehold and flexible stem regions (Fig. S1). When these conjugates are used to displace NP2 and NP3 from the circular scaffold, AuNP dimers connected by a length of double stranded DNA will be produced, and only one nanoparticle (NP1) will be left on the DNA circle (Fig. 1). To achieve the DNA nano-system, the first step is to generate a DNA circle. As shown in Fig. 2, during the formation of the DNA circle, strand C hybridized with CP (red region is complementary with the two ends of strand C). After hybridization, the binding of strand CP brought two ends of C into close proximity. Then, the gap was ligated by T4 ligase, and the circular DNA products were separated by 12% PAGE gel in TAE/Mg2+ buffer (Fig. 2, Lane 2). To confirm that the DNA circle was produced, Exonuclease I digestion was performed, since only ssDNA (single-stranded DNA) will be digested. The gel results suggested that strand CP was digested, leaving only DNA circles with quicker gel mobility (Fig. 2, lane 3). Likewise, strand displacement by adding strand CP-1 (completely complementary with strand CP) was also implemented to produce DNA circles (Fig. 2, lane 4). The newly observed DNA circle band indicated the separation of strand C and CP. Finally, digestion and displacement were both introduced. The gel results were similar to those of only adding Exonuclease I (Fig. 2, lane 5). The experimental results showed that either Exonuclease I digestion or strand displacement can effectively remove the strand used as template for ligation, releasing the DNA circle, which can then be purified by directly cutting it from the gel. The ligated DNA circles can now be used as a template for assembling DNA/AuNP conjugates by hybridization. In order to avoid formation of complicated network structures, we relied on monovalent recognition, i.e. one 15 nm diameter AuNP is connected only with one single thiola-DNA strand. In this way, aggregation induced by crosslinking of AuNPs functionalized with many AuNPs functionalized with many DNA strands will be avoided. Based on the designed AuNP assembly scheme, the additions of DNA and conjugates together led to a cooperative hybridization, thus resulting in the formation of different kinds of structures labeled A–E in Fig. 3a. Here, structure A was made by hybridization between a DNA/AuNP conjugate and a short single-stranded oligonucleotide NP2-DNA. Subsequently, structure B was formed by a DNA/AuNP conjugation with a DNA circle. For structure C,a pair
33
of AuNPs were connected by double-stranded DNA. Actually, the dimer structure was realized by complementary hybridization between two DNA/AuNP conjugates. Those AuNP assembly structures were constructed by mixing certain DNA/AuNP conjugates with DNA (linear or circular) for 8 h at room temperature. These gel electrophoresis experiments demonstrated that it was possible to separate the various AuNP/DNA assemblies. Firstly, the different AuNP structures were generated via hybridization. As shown in Fig. 3a, lanes 1–4 contained the products as types A, B, C and D, and showed the specific positions of the conjugations in the lanes. The sample loaded in lane 1 contained both NP1 and NP2 hybridized with a DNA circle, producing structure D (a pair of AuNPs attached to a single DNA circle). Although the products in lane 1 were mainly structure D, structure B, which was formed by the hybridization of a single NP and a DNA circle, was also observed. The reason for this phenomenon may be that NPs were not equivalently added into the hybridization reaction, and an excess amount of DNA circle was used. In lane 2, except for the band of structure D, an additional distinct band assigned to structure E was also observed, carrying groups of three particles, produced by hybridizing of NP1, 2, 3 and a DNA circle. The relatively low yield of E versus D can be explained as the aggregation of three gold particles having a stronger repulsive force which prevented them attaching to each other in contrast to the relatively easy aggregation of two particles. In lane 3, there was a band of structure C, which consisted of a pair of AuNPs structure (NP2 and AuDP2). In particular, a hybridized polymer of NP2 and AuDP2DNA was also loaded in lane 4, which indicated the products that would form upon addition of a displacement strand. To test the feasibility of our strategy, strand displacements were conducted as follows. In the presence of different displacing signals, variations in yield for the two kinds of displacing reactions (Fig. 1b and c) were observed in lanes 2, 3, 4 and lanes 5, 6, 7 in Fig. 3b. In the first case, the product of hybridizing by NP1, 2, 3 and DNA circle were displaced by AuDP2, AuDP3 separately and by both of them combined. Compared with products in lane 1 as a control, the newly produced band of structure C in lanes 2, 3, 4 (Fig. 3b) was a pair of AuNP products. However, there were still some unreacted initial AuNPs conjugates, which were groups of two and three particle polymers (D and E). In the second case, a displacing experiment was implemented by adding excess amount of AuDP2-DNA, AuDP3-DNA separately and both of them together. Interestingly, there was no structure C generated in lanes 5, 6, 7, but they had a weak band of structure B (which migrated slightly faster than C). Among them, the yield of B in lane 7 was larger than that in either lanes 5 or 6. Moreover, the band of initial AuNP
Fig. 2. Native PAGE (12%) analysis of the different formations of DNA assembly complexes. Lane 1: 20 bp DNA marker, lane 2: ligated DNA circle produced by hybridization of DNA strand C and CP, lane 3: ligated DNA circle digested by Exonuclease I, lane 4: ligated DNA circle displaced by linear complementary strand, lane 5: displaced products digested by Exonuclease I.
34
C. Zhang et al. / Journal of Colloid and Interface Science 418 (2014) 31–36
Fig. 3. Agarose gel electrophoretic analysis of DNA/AuNP conjugates. a Lanes: (1) polymer structure D, (2) polymer structure D and E, (3) a pair of AuNPs structure C, (4) hybridized polymer A. b (lane 1) Polymer structure D and E (lanes 2–4). The reactions displaced by AuDP2, AuDP3 separately and both of them combined, (lanes 5–7). The reactions implemented by adding AuDP2-DNA, AuDP3-DNA separately and the combination of both of them.
conjugates of E in lane 7 was rather weak, when compared with that of lanes 5 and 6. These phenomena resulted from the fact that, in each reaction, the DNA strand displacements really happened when adding specific DNA strands or DNA/AuNP conjugates. In the presence of either AuDP or AuDP-DNA, this nano-system can specifically react to different molecular signals, and the results clearly indicated the ability of strand displacement to alter the supramolecular assembly. However, the displacing effects of the two constructs were not the same. In Fig. 3b lane 7, the addition of AuDP-DNA led to nearly complete displacements with almost no D and E left. As a comparison, in lane 4, some of the polymers D and E still existed because of incomplete displacing. The possible reasons to explain these results were (1) the DNA toehold length required to trigger assembly of DNA/ AuNPs conjugates is longer (8 bp) than that of free DNA (5 bp) [30]; (2) although both AuDP and AuDP-DNA were added with excess amount, the concentrations of AuDP-DNA were higher than AuDP; (3) in the case of AuDP, the DNA strands were attached to AuNPs individually, which may inhibit the flexibility and movement of DNA strands during specific hybridizing recognition. Next, we took numerous TEM images of the products extracted from the gel bands to determine the structures of the band contents. As illustrated in Fig. 4, three kinds of DNA/AuNP conjugates were observed. For two-particle structures (C and D), it was hard to find the difference just from the TEM images, no matter whether
connected by dsDNA or circular ssDNA. The groups of three particles were clearly observed in the TEM images. To confirm the structures of specific bands in gel, the types of structures were counted in the TEM images. Here, we classified the observed structures into dimer, trimer, tetramer and aggregated derivatives. The statistical results are summarized in Fig. 4b. The results indicated that the production of structures D and E were not very high. This phenomenon may be caused by the incomplete band separation (intervals between adjacent bands were too narrow). In addition, the manual collection operations may be another reason for low production. From the results, the histogram of production of trimer E was about 29.2%, accompanied with high dimer production (14.8%). In contrast, for dimer D, the yield (35.8%) was much higher than that of trimer. Interestingly, there were two types of trimer structures: triangular and linear (Fig. 4). The possible reason for the formation of the linear triplet may be the rupture of the DNA circle, because it is difficult to have all connections using T4 ligase. In addition, the different arrangements of AuNPs may be another possible reason to explain this phenomenon (Supporting information). The percentage of observed triangular structures was 78.7%, higher than linear structures at 21.3%. This may be due to the strong repulsive force between three AuNPs, in which the triangular structure was a stable state when AuNPs conjugates hybridized on a DNA circle. These experimental results clearly indicated the ability to control the aggregations of DNA/AuNP conjugations.
C. Zhang et al. / Journal of Colloid and Interface Science 418 (2014) 31–36
35
Fig. 4. (a) Transmission electron micrographs of dimers and trimers: including structures of C, D and E. (b) Statistical results of TEM images. Percentage of production of polymer D (graph1) and E (graph2). The two structures of trimers were triangular and linear (graph3). The results are summarized in the histograms of the TEM images.
4. Conclusion In this work, a DNA/AuNP nanodevice has been developed, with the functions of assembling and releasing. Using this nanostructure device, single DNA/nanoparticle building blocks are capable to assemble into complex structures of two and three circular DNA/ AuNPs. The results show that this device can be controlled in a regulated fashion to yield expected discrete assemblies. However, compared with other complex DNA/AuNPs assembly structures, this study still has lots of room to improve [29]. In future research, the design should be considered to construct more complex catenanes DNA/AuNPs structures, combined with the methods of DNA origami and fluorescence. In addition, the circular DNA can keep its hybridizing ability without being digested by DNA enzymes, which may be used to form a DNA nanostructure via digesting enzymes. Further, since different DNA strands functionalized nanoparticles are released according to specific molecular signals, this DNA/AuNP nanodevice can provide a feasible tool to implement molecular targeting transport, and control the construction of a functional material step-by-step. More broadly, the ability and possibility of utilizing DNA strand migration to manipulate
complicated nano-systems are potentially demonstrated. The experimental results suggested that specific binding and displacing was the effective process for the flexible strand-induced molecular activities.
Acknowledgments This work was supported by the National Natural Science Foundation of China (61272161, 61127005, 61133010, 61033003, 60910002 and 61143003), Ph.D. Programs Foundation of the Ministry of Education of China (20110001130016), and Programme of Introducing Talents of Discipline to Universities (B13009), the program for Changjiang Scholars and Innovative Research Team in University (IRT0952). The Author would like to thank Prof. H. InakiSchlaberg for discussions and revisions.
Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jcis. 2013.11.079.
36
C. Zhang et al. / Journal of Colloid and Interface Science 418 (2014) 31–36
References [1] J. Sharma, R. Chhabra, A. Cheng, J. Brownell, Y. Liu, H. Yan, Science 323 (2009) 112–115. [2] N.D. Derr, B.S. Goodman, R. Jungmann, A.E. Leschziner, W.M. Shih, S.L. ReckPeterson, Science 338 (2012) 662–665. [3] P.K. Lo, P. Karam, F.A. Aldaye, C.K. McLaughlin, G.D. Hamblin, G. Cosa, H.F. Sleiman, Nat. Chem. 2 (2010) 319–323. [4] H. Gu, J. Chao, S. Xiao, N.C. Seeman, Nature 465 (2010) 202–206. [5] S.A. Claridge, A.J. Mastroianni, Y.B. Au, H.W. Liang, C.M. Micheel, J.M.J. Frechet, A.P. Alivisatos, J. Am. Chem. Soc. 130 (2008) 9598–9602. [6] A. Baeissa, N. Dave, B.D. Smith, J.W. Liu, ACS Appl. Mater. Interfaces 4 (2012) 1872–1877. [7] D.Y. Zhang, A.J. Turberfield, B. Yurke, E. Winfree, Science 318 (2007) 1121– 1124. [8] Y. Tian, C. Mao, J. Am. Chem. Soc. 126 (2004) 11410–11413. [9] F.A. Aldaye, H.F. Sleiman, J. Am. Chem. Soc. 129 (2007) 4130. [10] H. Pei, L. Liang, G.B. Yao, J. Li, Q. Huang, C.H. Fan, Angew. Chem. Int. Ed. 51 (2012) 9020–9024. [11] Y.Z. Xing, Z.Q. Yang, D.S. Liu, Angew. Chem. Int. Ed. 123 (2011) 12140–12142. [12] L.L. Qian, E. Winfree, J. Bruck, Nature 475 (2011) 368–372. [13] B. Chakraborty, R. Sha, N.C. Seeman, Proc. Natl. Acad. Sci. 105 (2008) 17245– 17249. [14] P. Yin, H.M.T. Choi, C.R. Calvert, N.A. Pierce, Nature 451 (2008) 318–323. [15] T. Omabegho, R. Sha, N.C. Seeman, Science 324 (2009) 67–71.
[16] A.J. Mastroianni, S.A. Claridge, A.P. Alivisatos, J. Am. Chem. Soc. 131 (2009) 8455–8499. [17] M.M. Maye, D. Nykypanchuk, M. Cuisinier, D. Van der Lelie, O. Gang, Nat. Mater. 8 (2009) 388–392. [18] M.M. Maye, M.T. Kumara, D. Nykypanchuk, W.B. Sherman, O. Gang, Nat. Nanotechnol. 5 (2010) 116–118. [19] B.Q. Ding, Z.T. Deng, H. Yan, S. Cabrini, R.N. Zuckermann, J. Bokor, J. Am. Chem. Soc. 132 (2010) 3248–3249. [20] A.M. Hung, C.M. Michee, L.D. Bozano, L.W. Osterbur, G.M. Wallraff, J.N. Cha, Nat. Nanotechnol. 5 (2010) 121–126. [21] J.Y. Kim, J.S. Lee, Nano Lett. 9 (2009) 4564–4569. [22] M.I. Shukoor, M.O. Altman, D. Han, A.T. Bayrac, I. Ocsoy, Z. Zhu, W.H. Tan, ACS Appl. Mater. Interfaces 4 (2012) 3007–3011. [23] X. Xu, N.L. Rosi, Y. Wang, F. Huo, C.A. Mirkin, J. Am. Chem. Soc. 128 (2006) 9286–9288. [24] S.A. Claridge, H.W. Liang, S.R. Basu, J.M.J. Frechet, A.P. Alivisatos, Nano Lett. 8 (2008) 1202–1205. [25] J.W. Kim, J.H. Kim, R. Deaton, Angew. Chem. Int. Ed. 50 (2011) 9185–9190. [26] N.H. Kim, S.J. Lee, M. Moskovits, Adv. Mater. 23 (2011) 4152–4156. [27] S.J. Barrow, A.M. Funston, D.E.G. Omez, T.J. Davis, P. Mulvaney, Nano Lett. 10 (2011) 4180–4187. [28] L.H. Guo, A.R. Ferhan, H. Chen, C.M. Li, G. Chen, S. Hong, D.H. Kim, Small 9 (2) (2013) 234–240. [29] A. Kuzyk, R. Schreiber, Z.Y. Fan, G. Pardatscher, E.M. Roller, A. Högele, F.C. Simmel, A.O. Govorov, T. Liedl, Nature 483 (2012) 311–314. [30] T.J. Song, H.J. Liang, J. Am. Chem. Soc. 134 (2012) 10803–10896.
Contents lists available at ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
Control of gold nanoparticles based on circular DNA strand displacement Cheng Zhang a, Jingjing Ma a,1, Jing Yang b, Yafei Dong c,⇑, Jin Xu a,⇑ a Institute of Software, School of Electronics Engineering and Computer Science, Key Laboratory of High Confidence Software Technologies of Ministry of Education, Peking University, No. 5 Yiheyuan Road, Haidian District, Beijing 100871, People’s Republic of China b School of Control and Computer Engineering, North China Electric Power University, No. 2 Beinong Road, Huilongguan, Changping District, Beijing 102206, People’s Republic of China c College of Life Science, Shannxi Normal University, 199 South Chang’an Road, Xi’an 710062, People’s Republic of China
a r t i c l e
i n f o
Article history: Received 6 September 2013 Accepted 28 November 2013 Available online 14 December 2013 Keywords: DNA strand displacement Circular DNA Gold nanoparticles DNA self-assembly TEM image
a b s t r a c t In this study, DNA strand displacement is utilized to control the aggregation of DNA/gold nanoparticles (AuNPs) based on circular DNA, in which DNA/AuNP conjugates are captured and released by adding different DNA signal strands. Using this strategy, single DNA/nanoparticle building blocks are capable of assembling into complex structures of two and three circular DNA/nanoparticles. The existence of these structures is demonstrated by gel electrophoresis and transmission electron microscopy (TEM) analysis. This advance has potential applications in controlling, transporting and detecting DNA/AuNP conjugates with subsequent manipulation of the structure and function of these assemblies. Ó 2013 Elsevier Inc. All rights reserved.
1. Introduction Programmable gold nanoparticle (AuNP) assembly leads to great advances in nano-construction. DNA-mediated self-assembly of AuNP is a promising way to organize and control a complicated nano-system. Through programmable DNA self-assembly, positions and geometries of AuNPs can be controlled precisely and hierarchically. A critical feature that contributes to the connection of DNA and AuNP is specific binding recognition (double-strand hybridization): that the attached DNA strands (DNA molecules) play a role as specific connection, which endow the homogenous AuNPs with different binding labels. A series of research works based on well-defined DNA/gold nanoparticle architectures have been established taking advantage of DNA base pairing [1–6]. Recently, flexible and deformable DNA nano-systems have been constructed, in which the method of strand displacement is used frequently, which is a mechanism of self-driving hybridization with DNA molecules branching migration [7–12]. By controlling several factors such as toehold sites (initial hybridizing region), strand lengths and hybridized structures, the DNA strands with the longer hybridizing regions can displace those with the shorter. In fact, the strand displacement method has been utilized in engineering many kinds of synthetic DNA machines [13–15]. It
⇑ Corresponding authors. E-mail addresses: [email protected] (C. Zhang), [email protected] (J. Ma), [email protected] (J. Yang), [email protected] (Y. Dong), DNA@ pku.edu.cn (J. Xu). 1 This author contributed equally with the first author. 0021-9797/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcis.2013.11.079
demonstrated that such biomolecular autocatalysis displacement can be driven by the free energy of specific base-pair (bp) hybridization, to make free energy of system to stable states. It has been a great challenge to fabricate reversible and complicated nanodevices based on gold nanoparticles. However, combined with DNA strand displacement, DNA/AuNP nanodevices capable of releasing and deforming functions have been developed [16–22]. In fact, many approaches for such applications have been proposed in recent years [21]. Weihong Tan’s group developed a Boolean logic operation in response to two pro-angiogenic targets in the area of cancer diagnosis and treatment [22]. In general, there are two main methods to prepare DNA/AuNP conjugates, both of which rely on thiolated DNA. One is functionalizing a single AuNPs with many DNA strands, which are randomly bound to the surface of particles [23]. Another is building AuNPs with a discrete number of single-stranded DNA [24]. Depending on different applications, numbers of oligonucleotides could be precisely connected on an individual nanoparticle [25]. Although prior work has advanced the field DNA/AuNP assembly, most examples that incorporated strand displacement capability were based on short linear DNA structures, thus limiting assemblies of more complicated hierarchical molecular scaffolds. In addition, many DNA/AuNP assemblies have been established choosing a strategy of settled structures, in which once the hybridizations complete, the positions and geometries of AuNPs will not change. Thus, the challenge of engineering nanodevices with tunable structures has attracted more interest [26–30]. In this study, we established a nanodevice based on DNA/AuNP conjugates displacement, in which DNA/AuNP conjugates were captured and
32
C. Zhang et al. / Journal of Colloid and Interface Science 418 (2014) 31–36
released selectively by adding different DNA signal molecules. The novel feature of our assemblies is the use of a circular DNA strand as a template on which to assemble DNA/AuNP conjugates, providing opportunities for construction of structurally more complex systems than has been possible before.
used filter paper to wick off excess buffer. After repeating washing for 1–2 times, grids were air-dried for analyzing (Supporting information Section 8).
2. Experimental materials and methods
As shown in Fig. 1, a nanodevice uses circular single-stranded DNA as a basic subunit, which consists of three regions, providing hybridization sites for DNA/AuNP conjugates NP1, NP2 and NP3. We also synthesized DNA/AuNP conjugates called AuDP2 and AuDP3, which are complementary to NP2 and NP3, and therefore are capable of displacing their respective partners from the circular DNA scaffold. After specific hybridization, three regions capture NP1, NP2 and NP3 by formation of double strand DNA lengths of 19 bp, 23 bp and 17 bp. Thus, three AuNPs are attached to one DNA circle to form a supramolecular assembly. In the design principle, the sequences of NP1, NP2 and NP3 all consist of three regions, including a flexible oligo-thymine domain to act as a spacer to extend the base pairing region away from the NP surface, a capturing domain complementary to a specific region of the circular template, and a 30 -terminal extension (Fig. S1). In addition, the DNA strands in NP2 and NP3 contained toehold domains between the spacer and recognition domains. These short sequences provide nucleation sites for the displacement strands used to disassemble the nanostructures. In the initial capture step, the DNA circle hybridizes with the complementary capture domain of the single DNA strands on DNA/AuNP conjugates NP1, NP2 and NP3. Then, to verify whether the DNA nanodevice can function as designed, a releasing course is implemented via strand displacement, in which toehold regions on NP conjugates will uniquely recognize single DNA strands on conjugates AuDP2 and AuDP3. Notably, a length of thymine bases is
2.1. Materials The main materials and chemicals used in our experiments are as follows: gold nanoparticles (AuNPs) with diameter of 15 nm from Ted Pella; Bis(p-sulfonatophenyl)phenylphosphinedipotassium salt dihydrate (BSPP) from Strem Chemical; Tris–Borate-EDTA buffer (TBE) and Tris–Acetic-EDTA buffer (TAE) from Solarbio; disulfide protected thiolated oligonucleotides by purification of HPLC from Sangon China; agarose from BIOWEST; methylenebisacrylamide and acrylamide monomer from TCI, Stains-All from SIGMA ALORICH. 2.1.1. Circularization of strand C DNA annealing (hybridization) and T4 ligase were used to form the DNA circle. The reactants strands C and CP (see Supporting information, Table S1 for sequences) were mixed together in TBE buffer and provided an annealing temperature conditions which were: 95 °C for 4 min, 65 °C for 30 min, 50 °C for 30 min, 37 °C for 30 min, 22 °C for 30 min. Then the solution was incubated with T4 ligase at 16 °C for 10 h. After ligation strand CP1 was added to hybridize with strand CP leaving the DNA circle free. 2.1.2. PAGE gel purification of DNA circle To separate DNA circle from other products non-denaturing polyacrylamide gel was used. The specific product of DNA circles were purified by directly cutting the target band and obtained by overnight alcohol precipitation at 20 °C (Suppporting information Section 3).
3. Results and discussion
2.1.3. Phosphination of AuNPs Solid powders of Bis(p-sulfonatophenyl) phenylphosphinedihydratedipotassium salt (BSPP) and AuNPs solution were mixed together and either shaken or stirred overnight at room temperature. Then NaCl (solid) was added slowly into the mixture. When the color of the solution changed from red to purple, it was centrifuged for 15 min and the liquid was removed. Finally, the AuNPs were dissolved in 1 BSPP solution. 2.1.4. Preparation of AuNPs-DNA conjugates DNA strands modified by disulfide were incubated with the AuNPs produced above for ligation (Suppporting information Section 6). Then 4% agarose gel was used to separate the AuNPsDNA conjugates. The AuNPs-DNA conjugates (NP and AuDP) were collected by using a centrifugal filter device. 2.1.5. Hybridization and displacement of AuNPs-DNA conjugates The collected NP conjugates were hybridized with DNA circles for 12 h at room temperature. Then AuDP conjugates were added into the solution for displacement of NP for 12 h at room temperature. The products were then electrophoresed in 4% agarose gels. 2.1.6. TEM images TEM images were obtained using a Tecnai G20. To increase surface hydrophilicity, TEM grids were ionized for 20–30 s using oxygen plasma cleaner (Harrick Plasma PDC-32). Then 2–4 lL of purified sample was spotted on the surface and left on the grid for 3 min. Later, added 10 lL of 0.5 TBE buffer for washing and
Fig. 1. Scheme of the assembly and release process. (a) Hybridizing to assemble into the three-particle structure. NP1, 2 and 3 indicate DNA/AuNP conjugates. (b) Displacement with DNA/AuNP conjugates. AuDP2 and 3 hybridize with NP2 and 3, leaving NP1 on the DNA circle. (c) Displacement with DNA strands, DNA/AuNP conjugates NP2 and 3 are removed from the DNA circle.
C. Zhang et al. / Journal of Colloid and Interface Science 418 (2014) 31–36
utilized to add sufficient flexibility and reduce straining of the structures, which would facilitate a flexible DNA strand structure turning during the displacement reaction. Similarly, at the far end of a single strand on the AuNP, another DNA region, also consisting of thymine bases, is designed for sequence length extension. In the same way, conjugates AuDP2 and AuDP3 also consist of hybridization, toehold and flexible stem regions (Fig. S1). When these conjugates are used to displace NP2 and NP3 from the circular scaffold, AuNP dimers connected by a length of double stranded DNA will be produced, and only one nanoparticle (NP1) will be left on the DNA circle (Fig. 1). To achieve the DNA nano-system, the first step is to generate a DNA circle. As shown in Fig. 2, during the formation of the DNA circle, strand C hybridized with CP (red region is complementary with the two ends of strand C). After hybridization, the binding of strand CP brought two ends of C into close proximity. Then, the gap was ligated by T4 ligase, and the circular DNA products were separated by 12% PAGE gel in TAE/Mg2+ buffer (Fig. 2, Lane 2). To confirm that the DNA circle was produced, Exonuclease I digestion was performed, since only ssDNA (single-stranded DNA) will be digested. The gel results suggested that strand CP was digested, leaving only DNA circles with quicker gel mobility (Fig. 2, lane 3). Likewise, strand displacement by adding strand CP-1 (completely complementary with strand CP) was also implemented to produce DNA circles (Fig. 2, lane 4). The newly observed DNA circle band indicated the separation of strand C and CP. Finally, digestion and displacement were both introduced. The gel results were similar to those of only adding Exonuclease I (Fig. 2, lane 5). The experimental results showed that either Exonuclease I digestion or strand displacement can effectively remove the strand used as template for ligation, releasing the DNA circle, which can then be purified by directly cutting it from the gel. The ligated DNA circles can now be used as a template for assembling DNA/AuNP conjugates by hybridization. In order to avoid formation of complicated network structures, we relied on monovalent recognition, i.e. one 15 nm diameter AuNP is connected only with one single thiola-DNA strand. In this way, aggregation induced by crosslinking of AuNPs functionalized with many AuNPs functionalized with many DNA strands will be avoided. Based on the designed AuNP assembly scheme, the additions of DNA and conjugates together led to a cooperative hybridization, thus resulting in the formation of different kinds of structures labeled A–E in Fig. 3a. Here, structure A was made by hybridization between a DNA/AuNP conjugate and a short single-stranded oligonucleotide NP2-DNA. Subsequently, structure B was formed by a DNA/AuNP conjugation with a DNA circle. For structure C,a pair
33
of AuNPs were connected by double-stranded DNA. Actually, the dimer structure was realized by complementary hybridization between two DNA/AuNP conjugates. Those AuNP assembly structures were constructed by mixing certain DNA/AuNP conjugates with DNA (linear or circular) for 8 h at room temperature. These gel electrophoresis experiments demonstrated that it was possible to separate the various AuNP/DNA assemblies. Firstly, the different AuNP structures were generated via hybridization. As shown in Fig. 3a, lanes 1–4 contained the products as types A, B, C and D, and showed the specific positions of the conjugations in the lanes. The sample loaded in lane 1 contained both NP1 and NP2 hybridized with a DNA circle, producing structure D (a pair of AuNPs attached to a single DNA circle). Although the products in lane 1 were mainly structure D, structure B, which was formed by the hybridization of a single NP and a DNA circle, was also observed. The reason for this phenomenon may be that NPs were not equivalently added into the hybridization reaction, and an excess amount of DNA circle was used. In lane 2, except for the band of structure D, an additional distinct band assigned to structure E was also observed, carrying groups of three particles, produced by hybridizing of NP1, 2, 3 and a DNA circle. The relatively low yield of E versus D can be explained as the aggregation of three gold particles having a stronger repulsive force which prevented them attaching to each other in contrast to the relatively easy aggregation of two particles. In lane 3, there was a band of structure C, which consisted of a pair of AuNPs structure (NP2 and AuDP2). In particular, a hybridized polymer of NP2 and AuDP2DNA was also loaded in lane 4, which indicated the products that would form upon addition of a displacement strand. To test the feasibility of our strategy, strand displacements were conducted as follows. In the presence of different displacing signals, variations in yield for the two kinds of displacing reactions (Fig. 1b and c) were observed in lanes 2, 3, 4 and lanes 5, 6, 7 in Fig. 3b. In the first case, the product of hybridizing by NP1, 2, 3 and DNA circle were displaced by AuDP2, AuDP3 separately and by both of them combined. Compared with products in lane 1 as a control, the newly produced band of structure C in lanes 2, 3, 4 (Fig. 3b) was a pair of AuNP products. However, there were still some unreacted initial AuNPs conjugates, which were groups of two and three particle polymers (D and E). In the second case, a displacing experiment was implemented by adding excess amount of AuDP2-DNA, AuDP3-DNA separately and both of them together. Interestingly, there was no structure C generated in lanes 5, 6, 7, but they had a weak band of structure B (which migrated slightly faster than C). Among them, the yield of B in lane 7 was larger than that in either lanes 5 or 6. Moreover, the band of initial AuNP
Fig. 2. Native PAGE (12%) analysis of the different formations of DNA assembly complexes. Lane 1: 20 bp DNA marker, lane 2: ligated DNA circle produced by hybridization of DNA strand C and CP, lane 3: ligated DNA circle digested by Exonuclease I, lane 4: ligated DNA circle displaced by linear complementary strand, lane 5: displaced products digested by Exonuclease I.
34
C. Zhang et al. / Journal of Colloid and Interface Science 418 (2014) 31–36
Fig. 3. Agarose gel electrophoretic analysis of DNA/AuNP conjugates. a Lanes: (1) polymer structure D, (2) polymer structure D and E, (3) a pair of AuNPs structure C, (4) hybridized polymer A. b (lane 1) Polymer structure D and E (lanes 2–4). The reactions displaced by AuDP2, AuDP3 separately and both of them combined, (lanes 5–7). The reactions implemented by adding AuDP2-DNA, AuDP3-DNA separately and the combination of both of them.
conjugates of E in lane 7 was rather weak, when compared with that of lanes 5 and 6. These phenomena resulted from the fact that, in each reaction, the DNA strand displacements really happened when adding specific DNA strands or DNA/AuNP conjugates. In the presence of either AuDP or AuDP-DNA, this nano-system can specifically react to different molecular signals, and the results clearly indicated the ability of strand displacement to alter the supramolecular assembly. However, the displacing effects of the two constructs were not the same. In Fig. 3b lane 7, the addition of AuDP-DNA led to nearly complete displacements with almost no D and E left. As a comparison, in lane 4, some of the polymers D and E still existed because of incomplete displacing. The possible reasons to explain these results were (1) the DNA toehold length required to trigger assembly of DNA/ AuNPs conjugates is longer (8 bp) than that of free DNA (5 bp) [30]; (2) although both AuDP and AuDP-DNA were added with excess amount, the concentrations of AuDP-DNA were higher than AuDP; (3) in the case of AuDP, the DNA strands were attached to AuNPs individually, which may inhibit the flexibility and movement of DNA strands during specific hybridizing recognition. Next, we took numerous TEM images of the products extracted from the gel bands to determine the structures of the band contents. As illustrated in Fig. 4, three kinds of DNA/AuNP conjugates were observed. For two-particle structures (C and D), it was hard to find the difference just from the TEM images, no matter whether
connected by dsDNA or circular ssDNA. The groups of three particles were clearly observed in the TEM images. To confirm the structures of specific bands in gel, the types of structures were counted in the TEM images. Here, we classified the observed structures into dimer, trimer, tetramer and aggregated derivatives. The statistical results are summarized in Fig. 4b. The results indicated that the production of structures D and E were not very high. This phenomenon may be caused by the incomplete band separation (intervals between adjacent bands were too narrow). In addition, the manual collection operations may be another reason for low production. From the results, the histogram of production of trimer E was about 29.2%, accompanied with high dimer production (14.8%). In contrast, for dimer D, the yield (35.8%) was much higher than that of trimer. Interestingly, there were two types of trimer structures: triangular and linear (Fig. 4). The possible reason for the formation of the linear triplet may be the rupture of the DNA circle, because it is difficult to have all connections using T4 ligase. In addition, the different arrangements of AuNPs may be another possible reason to explain this phenomenon (Supporting information). The percentage of observed triangular structures was 78.7%, higher than linear structures at 21.3%. This may be due to the strong repulsive force between three AuNPs, in which the triangular structure was a stable state when AuNPs conjugates hybridized on a DNA circle. These experimental results clearly indicated the ability to control the aggregations of DNA/AuNP conjugations.
C. Zhang et al. / Journal of Colloid and Interface Science 418 (2014) 31–36
35
Fig. 4. (a) Transmission electron micrographs of dimers and trimers: including structures of C, D and E. (b) Statistical results of TEM images. Percentage of production of polymer D (graph1) and E (graph2). The two structures of trimers were triangular and linear (graph3). The results are summarized in the histograms of the TEM images.
4. Conclusion In this work, a DNA/AuNP nanodevice has been developed, with the functions of assembling and releasing. Using this nanostructure device, single DNA/nanoparticle building blocks are capable to assemble into complex structures of two and three circular DNA/ AuNPs. The results show that this device can be controlled in a regulated fashion to yield expected discrete assemblies. However, compared with other complex DNA/AuNPs assembly structures, this study still has lots of room to improve [29]. In future research, the design should be considered to construct more complex catenanes DNA/AuNPs structures, combined with the methods of DNA origami and fluorescence. In addition, the circular DNA can keep its hybridizing ability without being digested by DNA enzymes, which may be used to form a DNA nanostructure via digesting enzymes. Further, since different DNA strands functionalized nanoparticles are released according to specific molecular signals, this DNA/AuNP nanodevice can provide a feasible tool to implement molecular targeting transport, and control the construction of a functional material step-by-step. More broadly, the ability and possibility of utilizing DNA strand migration to manipulate
complicated nano-systems are potentially demonstrated. The experimental results suggested that specific binding and displacing was the effective process for the flexible strand-induced molecular activities.
Acknowledgments This work was supported by the National Natural Science Foundation of China (61272161, 61127005, 61133010, 61033003, 60910002 and 61143003), Ph.D. Programs Foundation of the Ministry of Education of China (20110001130016), and Programme of Introducing Talents of Discipline to Universities (B13009), the program for Changjiang Scholars and Innovative Research Team in University (IRT0952). The Author would like to thank Prof. H. InakiSchlaberg for discussions and revisions.
Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jcis. 2013.11.079.
36
C. Zhang et al. / Journal of Colloid and Interface Science 418 (2014) 31–36
References [1] J. Sharma, R. Chhabra, A. Cheng, J. Brownell, Y. Liu, H. Yan, Science 323 (2009) 112–115. [2] N.D. Derr, B.S. Goodman, R. Jungmann, A.E. Leschziner, W.M. Shih, S.L. ReckPeterson, Science 338 (2012) 662–665. [3] P.K. Lo, P. Karam, F.A. Aldaye, C.K. McLaughlin, G.D. Hamblin, G. Cosa, H.F. Sleiman, Nat. Chem. 2 (2010) 319–323. [4] H. Gu, J. Chao, S. Xiao, N.C. Seeman, Nature 465 (2010) 202–206. [5] S.A. Claridge, A.J. Mastroianni, Y.B. Au, H.W. Liang, C.M. Micheel, J.M.J. Frechet, A.P. Alivisatos, J. Am. Chem. Soc. 130 (2008) 9598–9602. [6] A. Baeissa, N. Dave, B.D. Smith, J.W. Liu, ACS Appl. Mater. Interfaces 4 (2012) 1872–1877. [7] D.Y. Zhang, A.J. Turberfield, B. Yurke, E. Winfree, Science 318 (2007) 1121– 1124. [8] Y. Tian, C. Mao, J. Am. Chem. Soc. 126 (2004) 11410–11413. [9] F.A. Aldaye, H.F. Sleiman, J. Am. Chem. Soc. 129 (2007) 4130. [10] H. Pei, L. Liang, G.B. Yao, J. Li, Q. Huang, C.H. Fan, Angew. Chem. Int. Ed. 51 (2012) 9020–9024. [11] Y.Z. Xing, Z.Q. Yang, D.S. Liu, Angew. Chem. Int. Ed. 123 (2011) 12140–12142. [12] L.L. Qian, E. Winfree, J. Bruck, Nature 475 (2011) 368–372. [13] B. Chakraborty, R. Sha, N.C. Seeman, Proc. Natl. Acad. Sci. 105 (2008) 17245– 17249. [14] P. Yin, H.M.T. Choi, C.R. Calvert, N.A. Pierce, Nature 451 (2008) 318–323. [15] T. Omabegho, R. Sha, N.C. Seeman, Science 324 (2009) 67–71.
[16] A.J. Mastroianni, S.A. Claridge, A.P. Alivisatos, J. Am. Chem. Soc. 131 (2009) 8455–8499. [17] M.M. Maye, D. Nykypanchuk, M. Cuisinier, D. Van der Lelie, O. Gang, Nat. Mater. 8 (2009) 388–392. [18] M.M. Maye, M.T. Kumara, D. Nykypanchuk, W.B. Sherman, O. Gang, Nat. Nanotechnol. 5 (2010) 116–118. [19] B.Q. Ding, Z.T. Deng, H. Yan, S. Cabrini, R.N. Zuckermann, J. Bokor, J. Am. Chem. Soc. 132 (2010) 3248–3249. [20] A.M. Hung, C.M. Michee, L.D. Bozano, L.W. Osterbur, G.M. Wallraff, J.N. Cha, Nat. Nanotechnol. 5 (2010) 121–126. [21] J.Y. Kim, J.S. Lee, Nano Lett. 9 (2009) 4564–4569. [22] M.I. Shukoor, M.O. Altman, D. Han, A.T. Bayrac, I. Ocsoy, Z. Zhu, W.H. Tan, ACS Appl. Mater. Interfaces 4 (2012) 3007–3011. [23] X. Xu, N.L. Rosi, Y. Wang, F. Huo, C.A. Mirkin, J. Am. Chem. Soc. 128 (2006) 9286–9288. [24] S.A. Claridge, H.W. Liang, S.R. Basu, J.M.J. Frechet, A.P. Alivisatos, Nano Lett. 8 (2008) 1202–1205. [25] J.W. Kim, J.H. Kim, R. Deaton, Angew. Chem. Int. Ed. 50 (2011) 9185–9190. [26] N.H. Kim, S.J. Lee, M. Moskovits, Adv. Mater. 23 (2011) 4152–4156. [27] S.J. Barrow, A.M. Funston, D.E.G. Omez, T.J. Davis, P. Mulvaney, Nano Lett. 10 (2011) 4180–4187. [28] L.H. Guo, A.R. Ferhan, H. Chen, C.M. Li, G. Chen, S. Hong, D.H. Kim, Small 9 (2) (2013) 234–240. [29] A. Kuzyk, R. Schreiber, Z.Y. Fan, G. Pardatscher, E.M. Roller, A. Högele, F.C. Simmel, A.O. Govorov, T. Liedl, Nature 483 (2012) 311–314. [30] T.J. Song, H.J. Liang, J. Am. Chem. Soc. 134 (2012) 10803–10896.
