Junwye
Lee
a,
Rashid
Amin
a,
Byeonghoon
Kim
a,
Soyeon
Kim
a,
Chang-Won
Lee
b,
Jong Min
Kim
b,
Thomas H.
LaBean
c and
Sung Ha
Park
*a
aSungkyunkwan Advanced Institute of Nanotechnology (SAINT) and Department of Physics, Sungkyunkwan University, Suwon, 440-746, Korea. E-mail: sunghapark@skku.edu
bFrontier Research Laboratory, Samsung Advanced Institute of Technology (SAIT), Yongin, 446-712, Korea
cDepartments of Computer Science, Chemistry, and Biomedical Engineering, Duke University, Durham, NC 27708, USA
First published on 4th November 2011
We fabricated zigzag and folded DNA nanostructures by an angle control scheme. In order to give a solid verification of its operation, an open tube structure was also designed and it shows drastic dimensional changes compared with 2 dimensional zigzag and folded structures. These self-assembled artificial DNA structures would provide nanoscale-resolution templates for the alignment of various functional materials.
To demonstrate the ACS, double-crossover (DX) tiles6 were chosen as the model system. Normal DX tiles consist of two side-by-side duplexes linked by crossover junctions; these join to form a lattice through continuous linkages involving cohesive sticky ends with four neighboring DX tiles. The overall flatness of the DX lattice is maintained by the number of half turns of B-form DNA between the crossover junctions with 21 and 16 nucleotides giving ∼720° and ∼540° angles, respectively. According to our ACS, we can rotate one tile against the other by manipulating the number of nucleotides within the 21 nucleotide region between tiles (Fig. 1). When two neighboring DX tiles bind each other, the strand ends match such that the 5′ end of one strand on duplex B matches its rotational position with a 3′ end at the terminus of duplex C (Fig. 1a, black solid lines). If we intentionally add or remove a nucleotide at the end of duplex B, the position of the strand ends are rotated with an angle of 34.3° in the clockwise or counter-clockwise direction, respectively. For lattice formation, the DX2 tile should be rotated with such an angle to fit the DX1 terminus. For example, we can manipulate two nucleotides at the end of duplex B and cause the DX2 tile to rotate against the DX1 tile with an angle of 68.6°.
Fig. 1 DNA nanostructure design by an angle control scheme (ACS). (a) Isometric (left) and front views (right) of two adjacent double crossover (DX) tiles, DX1 and DX2. When a number of nucleotides is added or removed within the sticky-end region of duplex B, the adjacent DX2 tile is rotated by an angle of 34.3° × N in the clockwise, counter-clockwise direction, respectively. (b–f) Planar figures of various DNA nanostructures made by the ACS; Plane lattice (PL) (b), Zigzag lattices (ZLs) (c and d), Folded lattices (FLs) (e and f), respectively. The PL and ZLs are fabricated by continuous attachment between two different tiles, whereas the FLs consists of three tiles. The blue and red regions represent the nucleotide manipulated sites and side views of each DNA nanostructure are given beside the illustrations. |
The simplest ACS-based structures are a plane lattice (PL) and two different zigzag lattices (ZLs). In PL, which is the normal DX lattice, the structure remains planar because no nucleotides were manipulated at the ends of PL tiles (Fig. 1b). In contrast to PL, a nucleotide was manipulated at both ends of the Z1 tiles that are the components of ZL1. This causes the adjacent Z1-B tile to rotate in the clockwise or counter clockwise directions against the Z1-A with an angle of 34.3° along the duplex axis. Therefore, the side view of ZL1 represents a 2D zigzag form with an angle of 34.3° (Fig. 1c). In order to make a steeper shape, we designed the ZL2 with an angle of 68.6° by modifying two nucleotides at the ends of each tile (Fig. 1d). As mentioned above, in order to maintain the overall flatness of ZLs, the same number of nucleotides that were removed in the red regions were then added to the opposite sides of Z1 and Z2 tiles in the blue regions (Fig. 1c and d).
The applicability and versatility of the ACS can be seen in the folded lattices (FLs) that consist of three different tiles. For FL1, one nucleotide is removed from one side of a duplex of F1-A tile and another one is added to the diagonally opposite region. This diagonal manipulation on the other duplex of F1-A creates alternate angle changes on four binding sites with an angle of −34.3° and +34.3°, respectively (see red and blue regions in Fig. 1e). When the manipulated F1-A binds with F1-B and F1-C tiles, they form the repeating building units, and the FL1 is assembled by alternately attaching these building blocks (Fig. S4, ESI†). The same strategy was used for FL2 but it was designed with manipulations of two nucleotides on each side of the F2-A tiles. The side view of FL shows larger projections in directions orthogonal to the lattice plane compared with ZLs and we expected the AFM images of FLs to give distinct topographic features.
To confirm the topographical differences between these five different DNA nanostructure designs, we fabricated samples by annealing the synthetic oligonucleotides and characterizing by atomic force microscopy (AFM) using tapping mode under aqueous buffer. AFM images of each ACS-based structure exhibit large 2D lattice assemblies (Fig. 2). The inset images were obtained by noise-filtered Fourier transform reconstruction showing the periodicities of the lattices (Fig. 2 and Fig. S7, ESI†).9 Although the lattices appear to assemble as designed, we were unable to find sufficient evidence to verify the expected topographical differences within the ACS-based structures in the AFM images. Similar problems have been reported in other systems where AFM imaging of DNA nanostructures appear flatter than expected; this could be due to the strong interaction between the mica substrate and the negatively charged DNA backbone.8,10 Due to the strong Coulomb interactions, the ACS-based 2D lattices collapsed onto the substrate, making it difficult to distinguish between the various designs during height analysis. However, since we do indeed observe very large lattice pieces on the surface, we believe this indicates that the ACS-based 2D lattices likely exist as zigzag and folded sheets within the solution, according to the designs.
Fig. 2 AFM characterization of ACS-based DNA nanostructures. (a–e) Schematic diagrams and their corresponding AFM images of PL (a), ZL1 (b), ZL2 (c), FL1 (d), FL2 (e), respectively. The insets are noise-filtered reconstructed images by fast Fourier transform showing the periodicity of the crystals. Scale bars are 100 nm unless otherwise noted. |
In order to verify our understanding of the ACS designs, we built a design intended to form a 1D open tube (OT) structure. This design is somewhat similar to the bent-triple-crossover (BTX) tile structures described previously in the literature.11 In our case a two-tile system was adopted and the OT-A tile was manipulated in a manner similar to that used in the F1-A tile (Fig. 3a). The key point of the OT design that distinguishes it from the previous ACS-based 2D lattices is that there is no nucleotide compensation since we do not wish to maintain overall flatness for a 2D lattice. When the tiles bind each other, the missing nucleotides accumulate continuously such that they cannot form lattice but instead create a tube-like 1D structure. According to the design (Fig. 3b), there should be 9 OT tiles along the circumference of the OT tube leading to an offset of 10 bases leaving the structure as an opened tube structure. The 1D nature of the OT structure was easily observed by AFM (Fig. 3c). The AFM images show the drastically different morphology compared with the ACS-based 2D lattices. Consequently this result suggests that the ACS works properly in the all ACS structures.
Fig. 3 Open tube (OT) structure and its analyses. (a and b) Planar figure and schematic diagram of the OT structure with side views. In contrast with FL structures, there is no nucleotide compensation. This leads to an offset (∼3.4 nm) in both edges of the OT structure thus preventing the structure from forming closed tubes. (c) AFM image of the OT structure. AFM image exhibits dramatically different structures compared with 2D lattice nanostructures. This implies that the ACS design was successful in all the ACS-based DNA nanostructures. The scale bar in the main image is 5 μm. |
Footnote |
† Electronic supplementary information (ESI) available: Detailed materials, methods, and descriptions of synthesizing process of folded structures, DNA sequences, additional AFM images and their analyses are provided. See DOI: 10.1039/c1sm06379k |
This journal is © The Royal Society of Chemistry 2012 |