Molecular Precision at Micrometer Length Scales: Hierarchical Assembly of DNA−Protein Nanostructures Daniel Schiffels,*,†,‡ Veronika A. Szalai,† and J. Alexander Liddle*,† †
Center for Nanoscale Science and Technology, National Institute of Standards and Technology, Gaithersburg, Maryland 20899, United States ‡ Maryland NanoCenter, University of Maryland, College Park, Maryland 20742, United States S Supporting Information *
ABSTRACT: Robust self-assembly across length scales is a ubiquitous feature of biological systems but remains challenging for synthetic structures. Taking a cue from biologywhere disparate molecules work together to produce large, functional assemblieswe demonstrate how to engineer microscale structures with nanoscale features: Our selfassembly approach begins by using DNA polymerase to controllably create double-stranded DNA (dsDNA) sections on a single-stranded template. The single-stranded DNA (ssDNA) sections are then folded into a mechanically flexible skeleton by the origami method. This process simultaneously shapes the structure at the nanoscale and directs the large-scale geometry. The DNA skeleton guides the assembly of RecA protein filaments, which provides rigidity at the micrometer scale. We use our modular design strategy to assemble tetrahedral, rectangular, and linear shapes of defined dimensions. This method enables the robust construction of complex assemblies, greatly extending the range of DNA-based self-assembly methods. KEYWORDS: self-assembly, DNA origami, RecA, protein filament, DNA polymerase, yield
D
RecA is a 38 kDa protein from Escherichia coli with multiple functions related to homologous recombination and DNA repair.24 In vitro, RecA assembles into helical filaments on DNA in the presence of a nucleotide cofactor. Because the assembly of RecA onto dsDNA causes it to stretch and untwist, we incorporate flexible ssDNA linkages into our designs that can accommodate twist. Several properties make integration of these protein filaments with DNA-based self-assembly attractive: (i) assembly is fast (3 nm·s−1 to 10 nm·s−1)25 and can be controlled by the type and presence of nucleotide cofactor; filaments assembled with ATP and disassemble once ATP is hydrolyzed to ADP. Here, we exploit stable, irreversible filament assembly using the nonhydrolyzable ATP analogue ATP-γS. (ii) RecA filaments are mechanically more rigid than bare dsDNA: reported values of the persistence length, P (RecA on dsDNA with ATP-γS), are P = 630 nm26 and P = (962 ± 57) nm,27 providing similar rigidity to four-helix bundles created by DNA origami28 or DNA tiles.23 (iii) The number of base pairs (bp) in a dsDNA template defines the number of RecA monomers in a filament (1 RecA per 3 bp)29 and thus enables the formation of filaments with a precisely defined,
NA-based self-assembly is a powerful method with which to construct multifunctional, molecularly addressable nanostructures of arbitrary shape that can be used in applications ranging from single-molecule measurements to the fabrication of theranostic agents. Over the past three decades, DNA self-assembly has been used to create increasingly complex nanoscale structures. DNA crystals1 and nanotubes2,3 have been constructed by periodic assembly of DNA “tile” units. Finite-size, three-dimensional geometries have been realized using the “DNA origami” technique.4−7 In addition, innovations in routing,8−10 lengthening of the DNA scaffold strand,11 multistep self-assembly,12−14 staple sequences design,15 scaffold-free single-stranded tile assembly,16,17 and origami-seeded tile assembly18−20 have all increased the available design space. However, nanostructure fabrication based on DNA alone can suffer from low yields21,22 and is hampered by the need to strike a balance between size and mechanical rigidity.23 Despite recent efforts,22 typical assembly protocols, which employ large numbers of discrete components, offer little control over the assembly pathway, limiting the size, complexity, and yield of the structures that can be produced. Here, we show how a minimal amount of information encoded in a DNA template can be used to direct a two-stage, hierarchical selfassembly process, involving DNA polymerase, and the DNAbinding protein, RecA, to create otherwise inaccessible structures. © 2017 American Chemical Society
Received: January 15, 2017 Accepted: June 26, 2017 Published: June 26, 2017 6623
DOI: 10.1021/acsnano.7b00320 ACS Nano 2017, 11, 6623−6629
Article
www.acsnano.org
Article
ACS Nano
Figure 1. DNA origami skeleton design. (A) Scaffold strand routing path: color indicates the base index. (B) Skeleton strand placement on scaffold connection between helices 0 to 15. TEM images of tripods annealed without (C) and with (D) skeleton strands.
Figure 2. Tetrahedron formation by RecA protein filament assembly. Three-dimensional model of a DNA tripod before (A) and after (B) RecA assembly. (C) RecA filament model based on crystal structure from protein data bank entry 3CMX.29 (D) TEM image of RecA rigidified tetrahedron.
ponents and instead relying on generic components to construct a framework that supports the functional units.
programmed length. This is not possible for assembly on ssDNA30,31 (presumably due to secondary structure) or circular dsDNA (due to accumulation of DNA overtwist). Unfolded sections of scaffold strand in DNA origami have previously been employed as molecular springs to create prestressed tensegrity architectures.32 Building on the concept of unfolded DNA segments as mechanical components, here, we demonstrate how to transform ssDNA sections into mechanically rigid struts by DNA polymerase gap-filling and RecA protein assembly. Expanding the self-assembly toolbox by using the cell’s DNAhandling machinery, blending sequence-specific and structurespecific elements, enables us to make micrometer-scale, rigid, molecularly addressable structures that may be integrated with high-throughput, top-down fabrication methods. More generally, our results indicate that the scale of finite-size selfassembling systems can be dramatically increasedwithout affecting yieldby minimizing the number of unique com-
RESULTS AND DISCUSSION We first demonstrate the ability of RecA to rigidify a DNA structure by assembling a wireframe tetrahedron. We begin by defining a routing path for the scaffold strand (Figure 1A): first we choose three edges, connected by a single vertex, and construct them using the origami method, routing the scaffold strand six times through each edge to obtain rigid six-helixbundle nanotubes. The 177 distinct staple strands that fold the nanotubes are designed using caDNAno6 (Figure S1). Each pair of nanotubes has one three-base ssDNA connection at the vertex, conferring flexibility. Pairwise connections between the three remaining vertices at the tetrahedron base are made by 325 nucleotide ssDNA scaffold segments. To each unfolded scaffold segment, we add 10 “skeleton strands” (32-base deoxyoligonucleotides) programmed to form 6624
DOI: 10.1021/acsnano.7b00320 ACS Nano 2017, 11, 6623−6629
Article
ACS Nano a dsDNA segment (320 bp) with nine backbone nicks and a ssDNA gap (2 to 3 bases) bridging to the adjacent nanotube (Figure 1B). Annealing of staples and M13 scaffold without (Figure 1C) and with (Figure 1D) these skeleton strands yields tripod-shaped structures. Low-magnification transmission electron microscopy (TEM) images show that 85% of the structures are monomeric (not aggregated with other structures, Figure S2). High-magnification imaging of select multimeric structures reveals that multimer formation likely occurs from structures in intermediate folding states rather than association of fully formed tripods (Figure S3). The angles between the nanotube legs of the tripods vary over a wide range, consistent with the dsDNA segment length being a factor of 2 longer than its persistence length. Next, excess staples are removed and the magnesium concentration adjusted to 2 mM (Methods). Under this ionic condition, the tripod legs display more bends and have on average a smaller opening angle (Figure S4). RecA protein filament assembly on dsDNA segments yields rigid tetrahedra with well-defined edge lengths (Figure 2A,B). A 10 min incubation of tripods with 2-fold molar excess of RecA monomers to binding sites (exposed DNA nucleotide triplets) produces tetrahedra with three undamaged DNA origami edges and three RecA protein filament edges, which are clearly distinguishable by TEM (Figure 2D). We attribute the inability of RecA to form filaments on the DNA molecules within a nanotube to steric constraints imposed by the close-packed nanotube lattice: the RecA filament has a diameter ≈4.5 times larger than the spacing between DNA duplexes within the nanotube33 (Figure 2C). We estimate the assembly yield by shape analysis from TEM images (Figure S5). Among the N = 184 structures in the images, we identify 68 as “wellfolded” and another 112 that may either have been distorted during sample deposition (but were also “well-folded) or may have self-assembled with defects. Comparing our tetrahedra with those made by origami methods,34 we achieve a 1.6-fold increase in edge length without sacrificing mechanical rigidity. The scaffold strand defines the assembly pathway for the rigid RecA filaments efficiently, enabling the creation of larger structures without the use of the additional staples or larger scaffold strand that scale-up of conventional origami would require. We further reduce the number of input components via a strategy that obviates the need for skeleton strands. We demonstrate this on a rectangular structure in which skeleton strands make up 30% of the total number of DNA deoxyoligonucleotides. The structure comprises two origami nanotubes connected by two 1024 nucleotide long unfolded scaffold sections that serve as templates for RecA filaments (Figure S6). The target length is 522 nm (calculated based on an average nucleotide spacing of 0.51 nm).29 To replace the skeleton strands, we first use a polymerase “gap-fill” reaction to create defined regions of dsDNA on the ssDNA scaffold. Two polymerase start signals (primers) are placed on opposite sides of the M13 sequence. Two 3′-phosphorylated DNA molecules serve as stop signals (Figure 3A). The 960 nucleotide gap between primers and stoppers is filled by T7 DNA polymerase. The resulting partially doublestranded M13 DNA is then folded by staple strands. All assembly steps are monitored by a gel mobility assay (Figure 3B). Mobility before and after folding is consistent with controls in which the dsDNA sections are made using skeleton strands rather than polymerase. A faint band, visible in lane 2, can be explained by
Figure 3. DNA polymerase assembly and folding of partially double-stranded scaffold. (A) Primer (green) and stopper (red) placement for polymerase gap-fill reaction and model of DNA origami, folded with partially dsDNA scaffold. (B) Gel mobility assay. Control samples were prepared from unfolded M13 + skeleton staples (C1) and origami folded with skeleton staples (C2). (C) TEM image of origami, prepared using the polymerase scheme. (D) TEM image of origami folded from M13 with two dsDNA sections created by polymerase after the addition of RecA.
polymerase synthesis beyond a “stopper” point. This situation could occur by strand displacement or if the 3′-phosphate were missing from a “stopper”. Correct origami assembly with dsDNA connections is confirmed by TEM (Figures 3C and S7). A second approach, in which the origamis were folded first with the addition of two primers, relying on the origami nanotubes as stoppers, and the polymerase added second, resulted in structures that commonly displayed defects and was thus not used for RecA assembly experiments. It is possible that the polymerase’s nuclease activity destabilizes the origami structure. The addition of RecA to origamis, assembled using the polymerase-first scheme (Figure 3A), generates the desired rectangular structure (Figure 3D and Figures S8 and S9). We characterized protein filaments in these structures, as well as bare dsDNA connections in origamis before protein addition, by contour tracing and measurement of contour length, L, and end-to-end distance, R (Figure 4A,B). Complete DNA polymerase gap-filling creates 1024 bp long sections of dsDNA with an expected length of 348 nm, assuming 0.34 nm/bp for B-form DNA. The measured contour length (Figure 4C) is in excellent agreement, demonstrating the completeness of the gap-fill reaction. In RecA filaments, DNA is stretched to 0.51 nm/bp,29 resulting in an expected filament length of 522 nm. The measured contour length (Figure 4D) is again in excellent agreement, which shows complete RecA assembly. As 6625
DOI: 10.1021/acsnano.7b00320 ACS Nano 2017, 11, 6623−6629
Article
ACS Nano
(RDNA) to RecA filament connector end-to-end distances (RRecA) (Figure 4D). The mean of RRecA (421 nm) is over 3 times larger than that of RDNA (137 nm), with a narrower distribution. To measure persistence length, we further analyzed the tangent correlationthe average inner product between unit tangent vectorsas a function of tangent separation distance, Δx (see Methods). The tangent correlations decay exponentially (Figure 4E), and fitting eq 1 ⟨ t (⃗ x) t (⃗ x + Δx)⟩ = e−Δx / P
(1)
for 0 nm < Δx ≤ 100 nm, yields persistence lengths of PDNA = 54 nm and PRecA = 383 nm, demonstrating a significant increase in bending stiffness. Measurement uncertainties due to the finite sample size estimated using a bootstrap method23 (Supporting Information note S1 and Methods) are σBS = 3 nm and σBS = 26 nm for DNA and RecA, respectively. Effects of tracing accuracy, fit range, and surface−sample interactions are discussed in Supporting Information note S1. To demonstrate that our method is capable of creating stiff micrometer-scale structures beyond the size limitations of conventional origami, we designed a third structure, resembling a jump rope, folded from a linearized version of the M13 sequence, created using a restriction enzyme (Figures S10 and S11). Two nanotubes with distinct lengths and cross sections are connected by 3115 bp of dsDNA, created by polymerase (Figure 5A,B). The structure is transformed into a rigid filament by the addition of RecA (Figure 5C). Filament length is again tightly controlled, and the structure retains fully addressable origami on either end (Figure 5D).
CONCLUSIONS Our assembly approach is, in principle, less error-prone than other methods for creating large structures that increase size or rigidity by increasing the size of the scaffold and, concomitantly, the number of distinct staple strands needed.11 By decreasing the overall number of unique components, we reduce the intrinsic probability of an assembly error because we decrease the number of pathways along which a structure can misfold.21 Additionally, DNA synthesis by polymerase and RecA assembly are essentially error-free processes over the distances relevant to our method (≈1000 nucleotides).35 RecA is also small, diffusing rapidly, and binds nonsequence specifically, enabling fast assemblyunlike multiorigami assemblies.12−14,36 In the latter, both translational (∝ diameter−1) and rotational diffusion (∝ diameter−3) become slow, limiting the speed with which components can reach and orient to an attachment site, leading to assembly errors and low yields. Finally, the fact that RecA assembles into simple, linear structures and exhibits a high degree of cooperativity37 also serves to reduce the number of possible mis-assembly pathways. Our method is able to create large structures not only because of the larger persistence length conferred by RecA binding but also because much less DNA is required to template the RecA assembly than to create a similarly rigid structure using DNA alone: a comparably stiff four-helix bundle28 consumes about six times more scaffold nucleotides. The reduced assembly error rates of our approach make it scalable to longer scaffold strands, such as λ DNA. Use of longer scaffolds, in turn, enables the creation of structures with molecular precision and micrometer extent that could serve as interfaces for cells,38 in which we control both short-range and long-range spatial organization. At this size, they could also be
Figure 4. Mechanical deformation of DNA and RecA components. TEM images of origami with dsDNA (A) and RecA (B) connections with superimposed trace, indicating contour length, L, and end-toend distance, R. Inset: RecA section showing 10 manually selected intensity maxima used to count 4.5 helical turns. (C,D) Contour length and end-to-end distance histograms of dsDNA (red) and RecA (blue). Dashed lines show target contour length. (E) Tangent correlation plots with fit of eq 1.
an independent check for RecA coverage we measure the number of helical turns per filament (Figure 4B, inset). The average number of turns among 16 filaments is 56.2 with a standard deviation of 1.7, consistent with the number of 55 turns per filament calculated using the helical pitch of RecA derived from the crystal structure (1 turn/18.6 bp).29 Together these measurements establish that our three-step method (DNA polymerase gap-fill, origami folding, RecA filament assembly) creates well-defined protein filaments, enabling rational design of nanostructures. To quantify the enhancement of mechanical rigidity by RecA, we compared end-to-end distances, R, of dsDNA connectors 6626
DOI: 10.1021/acsnano.7b00320 ACS Nano 2017, 11, 6623−6629
Article
ACS Nano
Figure 5. Self-assembly of finite, micrometer-scale structures. (A) Placement of primer (green), stopper (red), and a second phosphorylated DNA oligonucleotide (blue), complementary to the BamHI recognition site, and 3D model of folded “jump rope” origami. (B,C) TEM images of folded origami before and after RecA assembly. (D) High-magnification image of the interface between Nanotube 1 and RecA filament. Using any of the three M13 concentrations yielded DNA origami tripods (after step 1) and tetrahedra (after step 3). Tripods for yield analysis (Figure S2) were made with a concentration of M13 of 10 nmol·L−1. Tetrahedra for tetrahedral yield analysis (Figure S5) were made with a concentration of M13 of 0.5 nmol·L−1. Rectangular Structure Assembly Protocol. 1. M13 (10 nmol·L−1), primer, and stopper sequences (100 nmol·L−1 each) were annealed in Tris-11 buffer from 80 to 20 °C over 30 min. 2. 1 μL of T7 DNA polymerase and 5 μL of dNTP were added to 50 μL of annealed M13 solution and incubated for 10 min at 20 °C. 3. M13 was purified using a column purification kit. M13 concentration after purification was estimated based on the amount of input M13 (in mol) and the output volume. 4. 10 μL of purified M13 was annealed in a reaction volume of 100 μL with 50 nmol·L−1 per staple strand in Tris-11 buffer in a DNA thermal cycler using a short high-temperature incubation (80 °C for 1 min), followed by a slow temperature ramp from 65 to 35 °C, over 999 min. 5. Staple strands were removed as described in protocol 1 step 2. 6. 5 μL of DNA origami, 9.5 μL of RecA, 5 μL of ATP-γS, and 180.5 μL of Tris-2 buffer were incubated at 37 °C for 25 min. Each DNA origami has 2048 bp of RecA template section, and thus the solution contains a ratio of RecA proteins to bp triplets of approximately 2:1. Control samples in Figure 3B with skeleton staples and no polymerase reaction were prepared as follows: C1: M13 (10 nmol·L−1) and skeleton strand sequences (100 nmol·L−1 each) were annealed in Tris-11 buffer from 80 to 20 °C over 30 min. C2: 5 nmol·L−1 of M13 and 50 nmol·L−1 per staple strand and 50 nmol·L−1 per skeleton strand were annealed in Tris-11 buffer in a DNA thermal cycler using a short high-temperature incubation (80 °C for 5 min), followed by a slow temperature ramp from 65 to 35 °C, over 999 min. “Jump Rope” Structure Assembly Protocol. 1. M13 (10 nmol·L−1), primer, and stopper sequences (100 nmol·L−1 each) and BamHI complement sequence (100 nmol·L−1) were annealed in Tris-11 buffer from 80 to 20 °C over 30 min. 2. 1 μL of BamHI was added per 50 μL and incubated at 37 °C for 90 min. 3. 1 μL of T7 DNA polymerase and 5 μL of dNTP were added to 50 μL of annealed M13 solution and incubated for 10 min at 20 °C. 4. M13 was purified using a column purification kit. M13 concentration after purification was estimated based on the amount of input M13 (in mol) and the output volume.
integrated into devices with micrometer-scale features patterned by conventional diffraction-limited optical lithography, enabling current origami placement approaches39 to cover large areas with high throughput. Our approach to the construction of large, rigid DNA structures maintains the programmability and addressability associated with DNA origami but overcomes many of the current limitations facing structural DNA nanotechnology. Perhaps more importantly, our work elucidates the essential principles that must be followed to generate large, high-yield, self-assembling structures using diffusion-mediated processes. Finally, we note that the presented design space for DNA nanostructures, employing both sequence- and structurespecific interactions, has been created using just two of the myriad proteins that exist in the cell’s DNA-handling armamentarium and could be combined with other strategies that employ protein−DNA interactions such as zinc-finger-mediated folding40 or sequence-specific placement of RecA−ssDNA complexes.41,42 We anticipate that using the full complement of tools that are already available for DNA biotechnology will enable the synthesis of more complex and multifunctional selfassembling systems.
METHODS All chemicals, buffers, and acronyms are listed in the Supporting Information Table S3. DNA origami design and DNA sequences are contained in the supplementary data. Tetrahedron Assembly Protocol. 1. M13, staple strands and skeleton strands were annealed in Tris-11 buffer in a DNA thermal cycler using a short hightemperature incubation (80 °C for 5 min), followed by a slow temperature ramp from 65 to 35 °C, over 999 min. The ratio of concentrations of skeleton strand to staple strand to M13 was 10:10:1. The absolute concentration of M13 was 10 nmol·L−1 (A), 5 nmol·L−1 (B), or 0.5 nmol·L−1 (C). 2. DNA origami solution was centrifuged in a 100 kDa molecular weight cut-off filter at 100 000 m·s−2 (10 000 g) for 10 min. Solution that passed the filter was removed; 490 μL of fresh Tris-2 buffer was added to the filter and the centrifugation repeated two more times. The concentration of DNA origami after filtration was estimated based on the amount of input DNA origami (in mol) and the output volume of approximately 10 μL. 3. Protein filaments were grown in Tris-2 buffer, at 37 °C for 20 min. We used concentrations of 0.6 nmol·L−1 DNA origami, 0.4 μmol·L−1 RecA, 0.2 mmol·L−1 ATP-γS. Each DNA origami has 975 bp of RecA template section and thus the solution contains a ratio of RecA proteins to bp triplets of approximately 2:1. 6627
DOI: 10.1021/acsnano.7b00320 ACS Nano 2017, 11, 6623−6629
Article
ACS Nano 5. 10 μL of purified M13 was annealed in a reaction volume of 100 μL with 50 nmol·L−1 per staple strand in Tris-11 buffer in a DNA thermal cycler using a short high temperature incubation (80 °C for 1 min), followed by a slow temperature ramp from 65 to 35 °C, over 999 min. 6. 2 μL of DNA origami, 11.6 μL of RecA, 5 μL of ATP-γS and 181.4 μL of Tris-2 buffer were incubated at 37 °C for 10 min. Each DNA origami has 3115 bp of RecA template section and thus the solution contains a ratio of RecA proteins to bp-triplets of approximately 2:1. Gel Electrophoresis. Agarose gels were prepared in TAE/Mg buffer with ethidium bromide and run in TAE/Mg buffer at 70 V for 90 to 120 min. Images were taken under UV light illumination. TEM Sample Preparation and Imaging. Carbon type-B TEM grids were placed in a plasma cleaner for 10 s. Two microliters of sample solution was deposited on a grid. After 2 min, the drop was removed using filter paper. Seven microliters of UAc solution was deposited on the sample and removed immediately followed by a second drop that was removed after 7 s. The grids were left to dry for at least 2 h before TEM imaging. Images of complete linear RecA structures (Figure 4C) were constructed from multiple, overlapping high-resolution images (typically four) that each captured one section of the assembly. Images were stitched using MosaicJ.43 Atomic Model Preparation. Atomic models of DNA origami were created using caDNAno6 and cando44 to create pdb files. A pdb file for RecA, assembled on dsDNA, was obtained from protein data bank entry 3CMX.29 Measurement of Contour Length, End-to-End Distance, and Persistence Length. Contours of dsDNA and RecA connectors in TEM images were traced manually using segmented lines with average segment lengths of 8.6 nm and 22.8 nm, respectively. Contour length was calculated as the sum of the length of all segments and end-to-end distance as the distance between the first and last trace point. To determine persistence length, we first fit a spline to each segmented line and determined trace coordinates along this fit with 1 nm spacing. We then computed the inner products of pairs of tangents to the fit located at positions x and x + Δx along the fit contour: t (⃗ x)· t (⃗ x + Δx). For each contour, for each Δx, the position, x, was moved along the contour in increments of Δx and the inner product of all tangent pairs (in all contours) obtained in this way was averaged: ⟨ t (⃗ x)· t (⃗ x + Δx)⟩. This tangent correlation is then plotted as a function of Δx (Figure 4E). For each Δx, an error bar is calculated from the standard deviation of tangent inner products divided by the square root of the number of tangents: std[ t (⃗ x)· t (⃗ x + Δx)]/ n , where n is the number of obtained tangent pairs. Persistence length, P, was obtained by fitting eq 1: y = e−Δx/P to the tangent correlation data for 0 nm < Δx ≤ 100 nm. The choice of eq 1, used for polymers diffusing in 3D instead of y = e−Δx/2P for polymers diffusing on a 2D surface, was motivated by the strong interaction of DNA with the TEM grid surface preventing contour equilibration after deposition. We thus assume polymer curvature on the length scale of 0 nm < Δx ≤ 100 nm to be preserved upon deposition, which has been suggested previously.45
AUTHOR INFORMATION Corresponding Authors
*E-mail: daniel.schiff
[email protected]. *E-mail:
[email protected]. ORCID
Daniel Schiffels: 0000-0002-6767-5606 J. Alexander Liddle: 0000-0002-2508-7910 Author Contributions
D.S. and J.A.L. conceived the project. D.S. performed experiments and data analysis. All authors contributed to data interpretation and manuscript preparation. Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS D.S. acknowledges support under the Cooperative Research Agreement between the University of Maryland and the National Institute of Standards & Technology Center for Nanoscale Science and Technology, Award Number 70ANB10H193, through the University of Maryland. We thank Samuel Stavis, Mike Zwolak, and Danielle Schultz for helpful discussions. REFERENCES (1) Winfree, E.; Liu, F.; Wenzler, L. A.; Seeman, N. C. Design and Self-Assembly of Two-Dimensional DNA Crystals. Nature 1998, 394, 539−544. (2) Rothemund, P. W.; Ekani-Nkodo, A.; Papadakis, N.; Kumar, A.; Fygenson, D. K.; Winfree, E. Design and Characterization of Programmable DNA Nanotubes. J. Am. Chem. Soc. 2004, 126, 16344−16352. (3) Yin, P.; Hariadi, R. F.; Sahu, S.; Choi, H. M.; Park, S. H.; Labean, T. H.; Reif, J. H. Programming DNA Tube Circumferences. Science 2008, 321, 824−826. (4) Rothemund, P. W. Folding DNA to Create Nanoscale Shapes and Patterns. Nature 2006, 440, 297−302. (5) Douglas, S. M.; Dietz, H.; Liedl, T.; Hogberg, B.; Graf, F.; Shih, W. M. Self-Assembly of DNA into Nanoscale Three-Dimensional Shapes. Nature 2009, 459, 414−418. (6) Douglas, S. M.; Marblestone, A. H.; Teerapittayanon, S.; Vazquez, A.; Church, G. M.; Shih, W. M. Rapid Prototyping of 3D DNA-Origami Shapes with caDNAno. Nucleic Acids Res. 2009, 37, 5001−5006. (7) Dietz, H.; Douglas, S. M.; Shih, W. M. Folding DNA into Twisted and Curved Nanoscale Shapes. Science 2009, 325, 725−730. (8) Han, D.; Pal, S.; Nangreave, J.; Deng, Z.; Liu, Y.; Yan, H. DNA Origami with Complex Curvatures in Three-Dimensional Space. Science 2011, 332, 342−346. (9) Zhang, F.; Jiang, S.; Wu, S.; Li, Y.; Mao, C.; Liu, Y.; Yan, H. Complex Wireframe DNA Origami Nanostructures with Multi-Arm Junction Vertices. Nat. Nanotechnol. 2015, 10, 779−784. (10) Benson, E.; Mohammed, A.; Gardell, J.; Masich, S.; Czeizler, E.; Orponen, P.; Hogberg, B. DNA Rendering of Polyhedral Meshes at the Nanoscale. Nature 2015, 523, 441−444. (11) Marchi, A. N.; Saaem, I.; Vogen, B. N.; Brown, S.; LaBean, T. H. Toward Larger DNA Origami. Nano Lett. 2014, 14, 5740−5747. (12) Liu, W.; Zhong, H.; Wang, R.; Seeman, N. C. Crystalline TwoDimensional DNA-Origami Arrays. Angew. Chem., Int. Ed. 2011, 50, 264−267. (13) Rajendran, A.; Endo, M.; Katsuda, Y.; Hidaka, K.; Sugiyama, H. Programmed Two-Dimensional Self-Assembly of Multiple DNA Origami Jigsaw Pieces. ACS Nano 2011, 5, 665−671. (14) Iinuma, R.; Ke, Y.; Jungmann, R.; Schlichthaerle, T.; Woehrstein, J. B.; Yin, P. Polyhedra Self-Assembled from DNA
ASSOCIATED CONTENT S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsnano.7b00320. Figures S1−S11, Tables S1−S7, and note S1 with detailed descriptions of DNA origami design, assembly yield analysis, characterization of multimeric structures, additional low-magnification and high-magnification TEM images, measurement of RecA filament length distribution, reagent list, and a list of all DNA sequences (PDF) 6628
DOI: 10.1021/acsnano.7b00320 ACS Nano 2017, 11, 6623−6629
Article
ACS Nano Tripods and Characterized with 3D DNA-PAINT. Science 2014, 344, 65−69. (15) Niekamp, S.; Blumer, K.; Nafisi, P. M.; Tsui, K.; Garbutt, J.; Douglas, S. M. Folding Complex DNA Nanostructures from Limited Sets of Reusable Sequences. Nucleic Acids Res. 2016, 44, e102. (16) Wei, B.; Dai, M.; Yin, P. Complex Shapes Self-Assembled from Single-Stranded DNA Tiles. Nature 2012, 485, 623−626. (17) Ke, Y.; Ong, L. L.; Shih, W. M.; Yin, P. Three-Dimensional Structures Self-Assembled from DNA Bricks. Science 2012, 338, 1177− 1183. (18) Mohammed, A. M.; Schulman, R. Directing Self-Assembly of DNA Nanotubes Using Programmable Seeds. Nano Lett. 2013, 13, 4006−4013. (19) Mohammed, A. M.; Velazquez, L.; Chisenhall, A.; Schiffels, D.; Fygenson, D. K.; Schulman, R. Self-Assembly of Precisely Defined DNA Nanotube Superstructures Using DNA Origami Seeds. Nanoscale 2017, 9, 522−526. (20) Mohammed, A. M.; Sulc, P.; Zenk, J.; Schulman, R. SelfAssembling DNA Nanotubes to Connect Molecular Landmarks. Nat. Nanotechnol. 2016, 12, 312−316. (21) Murugan, A.; Zou, J.; Brenner, M. P. Undesired Usage and the Robust Self-Assembly of Heterogeneous Structures. Nat. Commun. 2015, 6, 6203. (22) Dunn, K. E.; Dannenberg, F.; Ouldridge, T. E.; Kwiatkowska, M.; Turberfield, A. J.; Bath, J. Guiding the Folding Pathway of DNA Origami. Nature 2015, 525, 82−86. (23) Schiffels, D.; Liedl, T.; Fygenson, D. K. Nanoscale Structure and Microscale Stiffness of DNA Nanotubes. ACS Nano 2013, 7, 6700− 6710. (24) Cox, M. M. Motoring Along with the Bacterial RecA Protein. Nat. Rev. Mol. Cell Biol. 2007, 8, 127−138. (25) Galletto, R.; Amitani, I.; Baskin, R. J.; Kowalczykowski, S. C. Direct Observation of Individual RecA Filaments Assembling on Single DNA Molecules. Nature 2006, 443, 875−878. (26) Egelman, E. H.; Stasiak, A. Structure of Helical RecA-DNA Complexes. Complexes Formed in the Presence of ATP-Gamma-S or ATP. J. Mol. Biol. 1986, 191, 677−697. (27) Hegner, M.; Smith, S. B.; Bustamante, C. Polymerization and Mechanical Properties of Single RecA-DNA Filaments. Proc. Natl. Acad. Sci. U. S. A. 1999, 96, 10109−10114. (28) Kauert, D. J.; Kurth, T.; Liedl, T.; Seidel, R. Direct Mechanical Measurements Reveal the Material Properties of Three-Dimensional DNA Origami. Nano Lett. 2011, 11, 5558−5563. (29) Chen, Z.; Yang, H.; Pavletich, N. P. Mechanism of Homologous Recombination from the RecA-ssDNA/dsDNA Structures. Nature 2008, 453, 489−484. (30) Flory, J.; Radding, C. M. Visualization of RecA Protein and its Association with DNA: A Priming Effect of Single-Strand-Binding Protein. Cell 1982, 28, 747−756. (31) Thresher, R. J.; Christiansen, G.; Griffith, J. D. Assembly of Presynaptic Filaments. Factors Affecting the Assembly of RecA Protein onto Single-Stranded DNA. J. Mol. Biol. 1988, 201, 101−113. (32) Liedl, T.; Hogberg, B.; Tytell, J.; Ingber, D. E.; Shih, W. M. SelfAssembly of Three-Dimensional Prestressed Tensegrity Structures from DNA. Nat. Nanotechnol. 2010, 5, 520−524. (33) Fischer, S.; Hartl, C.; Frank, K.; Radler, J. O.; Liedl, T.; Nickel, B. Shape and Interhelical Spacing of DNA Origami Nanostructures Studied by Small-Angle X-ray Scattering. Nano Lett. 2016, 16, 4282− 4287. (34) Smith, D. M.; Schuller, V.; Forthmann, C.; Schreiber, R.; Tinnefeld, P.; Liedl, T. A Structurally Variable Hinged Tetrahedron Framework from DNA Origami. J. Nucleic Acids 2011, 2011, 360954. (35) Kunkel, T. A. DNA Replication Fidelity. J. Biol. Chem. 2004, 279, 16895−16898. (36) Tikhomirov, G.; Petersen, P.; Qian, L. Programmable Disorder in Random DNA Tilings. Nat. Nanotechnol. 2016, 12, 251−259. (37) Stasiak, A.; Di Capua, E.; Koller, T. Elongation of Duplex DNA by RecA Protein. J. Mol. Biol. 1981, 151, 557−564.
(38) Angelin, A.; Weigel, S.; Garrecht, R.; Meyer, R.; Bauer, J.; Kumar, R. K.; Hirtz, M.; Niemeyer, C. M. Multiscale Origami Structures as Interface for Cells. Angew. Chem., Int. Ed. 2015, 54, 15813−15817. (39) Gopinath, A.; Miyazono, E.; Faraon, A.; Rothemund, P. W. Engineering and Mapping Nanocavity Emission via Precision Placement of DNA Origami. Nature 2016, 535, 401−405. (40) Praetorius, F.; Dietz, H. Self-Assembly of Genetically Encoded DNA-Protein Hybrid Nanoscale Shapes. Science 2017, 355, eaam5488. (41) Sharma, R.; Davies, A. G.; Walti, C. Nanoscale Programmable Sequence-Specific Patterning of DNA Scaffolds Using RecA Protein. Nanotechnology 2012, 23, 365301. (42) Sharma, R.; Davies, A. G.; Walti, C. Directed Assembly of 3-nmLong RecA Nucleoprotein Filaments on Double-Stranded DNA with Nanometer Resolution. ACS Nano 2014, 8, 3322−3330. (43) Thevenaz, P.; Unser, M. User-Friendly Semiautomated Assembly of Accurate Image Mosaics in Microscopy. Microsc. Res. Tech. 2007, 70, 135−146. (44) Pan, K.; Kim, D. N.; Zhang, F.; Adendorff, M. R.; Yan, H.; Bathe, M. Lattice-Free Prediction of Three-Dimensional Structure of Programmed DNA Assemblies. Nat. Commun. 2014, 5, 5578. (45) Abels, J. A.; Moreno-Herrero, F.; van der Heijden, T.; Dekker, C.; Dekker, N. H. Single-Molecule Measurements of the Persistence Length of Double-Stranded RNA. Biophys. J. 2005, 88, 2737−2744.
6629
DOI: 10.1021/acsnano.7b00320 ACS Nano 2017, 11, 6623−6629