Perspective pubs.acs.org/JACS
Cite This: J. Am. Chem. Soc. 2018, 140, 17808−17819
Concept and Development of Framework Nucleic Acids Zhilei Ge,†,∥ Hongzhou Gu,‡,∥ Qian Li,† and Chunhai Fan*,†,§ †
Downloaded via UNIV OF NEW ENGLAND on January 13, 2019 at 13:36:51 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
School of Chemistry and Chemical Engineering, and Institute of Molecular Medicine, Renji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai 200240, China ‡ Center for Biotechnology and Biomedical Engineering, Institutes of Biomedical Sciences, Fudan University, Shanghai 200032, China § Division of Physical Biology & Bioimaging Center, Shanghai Synchrotron Radiation Facility, Shanghai Institute of Applied Physics, Chinese Academy of Sciences, Shanghai 201800, China (1D). In Seeman’s original vision, DNA junctions with N (N ≥ 3) arms could be assembled in the same way, and with designed sticky ends for hybridization, the 6-arm junction might be able to further self-assemble into a cubic lattice, which could potentially serve as a skeleton frame to guide and orient protein molecules through DNA−protein interaction to enable protein crystallization.9 This vision opened up the era of structural DNA nanotechnology. Although the latter part of the vision using DNA framework to solve protein crystallizationis debated, a variety of framework-like DNA nanostructures, e.g., polyhedra,12,17−19 nanotubes,20−22 2D arrays,23−26 etc., have been created to precisely organize molecules, including inorganic nanoparticles, nucleic acids, lipids, peptides, and proteins, for material and medical sciences during the growing and branching process of the field.27−34 Not surprisingly, the design principle also applies to the creation of framework DNA-RNA hybridized or pure RNA nanostructures.35−37 Based on the features of spatial organization with molecular precision and superior programmability, many DNA/RNA nanostructures have found their applications in molecular and cellular biophysics, photonic and catalytic chemistry, diagnostics, as well as therapeutics. Meanwhile, as the configuration of DNA nanostructures goes from simple to complex, a transition of computer-aided programs from basic to advanced levels has occurred, which greatly facilitates the design of DNA nanostructures and urges the development of the field. The templates out of structures are imperative for constructing materials with different sizes on various scales (Figure 2a). For example, on the nanoscale, the construction of modular framework leads to the production of numerous integrated and functional architectures. Nature has harnessed this principle to evolve structural proteins to support and maintain physiological functions of cells, as exemplified by cytoskeleton proteins that form scaffolds to maintain the cell shape. In contrast, the scaffolding functions of the other two types of biomolecules involved in the Central Dogma, DNA and RNA, are less explored in biology, which have nevertheless been the focus of study in structural DNA nanotechnology. Especially, the development of shell or skeleton DNA frameworks, or more recently dubbed “framework nucleic acids” (FNAs) sheds new light on spatial organization of molecules and materials in vitro and in vivo.38−40
ABSTRACT: The blooming field of structural DNA nanotechnology harnessing the material properties of nucleic acids has attracted widespread interest. The exploitation of the precise and programmable Watson− Crick base pairing of DNA or RNA has led to the development of exquisite nucleic acid nanostructures from one to three dimensions. The advances of computer-aided tools facilitate automated design of DNA nanostructures with various sizes and shapes. Especially, the construction of shell or skeleton DNA frameworks, or more recently dubbed “framework nucleic acids” (FNAs) provides a means to organize molecules or nanoparticles with nanometer precision. The intrinsic biological properties and tailorable functionalities of FNAs hold great promise for physical, chemical, and biological applications. This Perspective highlights state-of-the-art design and construction, of precisely assembled FNAs, and outlines the challenges and opportunities for exploiting the structural potential of FNAs for translational applications.
1. INTRODUCTION The double helical structure model of DNA forms the basis of genetics ranging from microorganisms, plants to animals.1,2 Besides this best-known structure, nucleic acids have shown great versatility in nature. For example, single- or doublestranded riboswitches, ribozymes,3 circular RNA,4 and more complex four-stranded G-quadruplex5,6 as well as I-motif7,8 (Figure 1a) have been found to exist and play important regulatory roles in living cells. In test tubes, researchers have created even more complex multistranded DNA structures, including DNA tiles,9−11 DNA tetrahedron,12−14 DNA origami,15 as well as hybridized DNA-inorganic spherical nucleic acids16 (SNAs) (Figure 1b). These natural and artificial DNA structures offer high structural potential for engineering intracellular and in vivo biological processes. The B-form double-stranded (ds) DNA is a perfect natural polymer in the nanoscale. Its helix owns a diameter of ∼2 nm and a helical turn of ∼3.4 nm (or 10.4 bp). In the 1980s, Seeman demonstrated the in vitro assembly of an immobile Holliday (4-arm) junction, by stoichiometrically mixing four synthetic single-stranded (ss) DNAs with defined sequences to break the sequence symmetry of the mobile Holliday junction in cells.9,10 The success of the artificial 4-arm DNA junction turned DNA into a nanoscale polymer beyond one dimension © 2018 American Chemical Society
Received: September 29, 2018 Published: December 5, 2018 17808
DOI: 10.1021/jacs.8b10529 J. Am. Chem. Soc. 2018, 140, 17808−17819
Perspective
Journal of the American Chemical Society
Figure 1. Structural versatility of nucleic acids. (a) Structures that exist in nature. (b) Structures that are artificially designed. From left to right, trends are shown as the increasing number of strands to generate the structure. Images reproduced with permission: ref 4, Copyright 1988 Nature Publishing Group; ref 6, Copyright 2018 Nature Publishing Group; ref 11, Copyright 1991 American Chemical Society; ref 14, Copyright 2009 American Chemical Society; ref 15, Copyright 2006 Nature Publishing Group; ref 3 and 16, Copyright 2012 American Chemical Society.
of sequences that maximize the likelihood of desired sequence interactions while minimizing the likelihood of undesired interactions. The former was set to minimize sequence symmetry, by treating short sequences as vocabulary elements and limiting the number of repetitive elements, to gain control over the resulting secondary structure. The latter not only included the paradigm of sequence symmetry minimization but also incorporated the paradigm of energy minimization to maximize the strength of desired interactions. With the two programs, a series of polygonal DNA nanostructures17−19,53 were successfully designed and constructed, including tetrahedron53one of the most popular FNAs that have been widely used for biosensing13,38,54−66 and biophysics research.67,68 However, the laborious and inefficient preparation process limited the development of the nascent field. Meanwhile, the complexity of the so-prepared FNAs was restricted due to relatively simple geometric shapes or repetition of the basic tiles assembled from the component strands. In 2006, the invention of the DNA origami approach15 resolved most of the limitations in the design and construction of 2D DNA nanostructures: To design a desired shape, the 7kilobase M13 genomic ssDNA was raster filled into the frame shape as a scaffold, and hundreds of short synthetic staple strands were programmed to hold the scaffold in place; in the construction process, sequence optimization, purification, and strict stoichiometry of the component strands were all unnecessary. The size and complexity of DNA nanostructures encountered an explosion as the origami was invented.69,70 To model large and complex origami-based DNA nanostructures efficiently, the 3D graphic tools such as Tiamat71 and
FNAs represent a new type of nucleic acids with unique physical, chemical, and biological properties. First, FNAs are truly monodispersed nanostructures with near-atomistic precision, which compares favorably with less precise inorganic, organic, or polymeric nanoparticles (Figure 2b). Second, FNAs enable site-specific organization of small or macromolecules, or nanoparticles with nanometer precision.41−45 Third, FNAs demonstrate striking difference with single-stranded (ss-) or double-stranded (ds-) DNA for cellular internalization and biodistribution in vivo.46−50 By taking mechanical properties of FNAs as an example, we reason that the stiffness of FNAs are highly tunable and programmable (Figure 2c). Given that the persistence length of a polymer is proportional to the fourth order of the radius, the design of a FNA with the radius in the range of 10−100 nm leads to a nanostructure with a persistence length exceeding those of ss/ ds DNA by 4−6 orders of magnitude (ssDNA, 1 nm; and dsDNA, 50 nm). Hence, seamless interfacing of flexible ssDNA handles to rigid FNA scaffolds provides a designer route to constructing architectures on the nanoscale. In this Perspective, we focus on recent breakthroughs in the design of FNAs, and highlight the exploration of FNAs for intelligent membrane engineering, biosensing, imaging and drug delivery.
2. ORIGAMI AND AUTO-CAD TOOLS: LEAPS FOR FNA DEVELOPMENT Before 2000, DNA nanostructures were usually designed via a process to optimize the sequences of the component strands, and then constructed by mixing the purified strands in a strict stoichiometric ratio for self-assembly. The programs SEQUIN51 and NANEV52 were frequently used for the design 17809
DOI: 10.1021/jacs.8b10529 J. Am. Chem. Soc. 2018, 140, 17808−17819
Perspective
Journal of the American Chemical Society
(cadnano.org), a beginner may be able to complete his/her debuting design in hours. Besides, in the new version (caDNAno2) the tool has been made as a plugin for Autodesk Maya and can now display interactive 3D geometry in addition to the standalone views. The powerfulness of Maya in 3D picturing and animation further allows for the precise sketching and showcase of the origami-based FNAs with the guest molecules organized on them. The 3D modeling in caDNAno can predict the 3D structure of DNA origami designs with simple shapes, not including those with twist, stretch, or bend in customized DNA nanostructures. Given the substantial time and financial cost associated with the design and validation, CanDo74,75 was developed as a more professional modeling tool to predict 3D DNA structure, especially for the designs with combinations of twist-stretch−bend, before initiating cost-intensive staple strands synthesis. The tool accepts both Tiamat and caDNAno design files, and accounts for the canonical twist, bend and stretch stiffness of double-helical DNA domains, as well as nicks in the DNA double helix, entropic elasticity of ssDNA, and distant crossovers required to model wireframe DNA nanostructures.74,75 CanDo predicts not only the 3D solution shape of complex DNA nanostructures but also their mechanical flexibility, which significantly increases the number and variety of DNA nanostructures. The current tools Tiamat and caDNAno undergo the bottom-up pathway in the design process, wherein the specific Watson−Crick base pairing is manually designed. This requires the user grasping certain level of knowledge of the field prior to the design, and thus to some extent is unfriendly to outsiders. In addition, for the design of complex wireframe origamis, a lot of manual adjustment is required, making it very difficult to realize even for DNA nanotechnologists with expertise. These issues expedited the development of automated design procedures (Figure 3b).73,76 In one example, using routing algorithms based on graph theory and relaxation simulations that trace scaffold strands through the target structures, the design process can go top-down as follows: (1) Draw a desired shape in 3D software like Autodesk Maya, (2) Generate an appropriate routing of the long scaffold strand and relax the resulting DNA helix arrangement with scripts, leading to the wireframe approximation of the 3D shape, (3) Fine-tune the design and generate the staple strands with Autodesk Maya running vHelix, a custom-made dedicated plugin for the design and visualization of DNA nanostructures.76 In this procedure, a certain level of manual adjustment is still necessary. An alternative procedure named DAEDALUS achieves the full automation without reliance on user feedback.73 Both procedures have been demonstrated to facilitate the rendering of arbitrary polyhedral DNA nanostructures. They also open up the possibility of “one-click” 3D printing of FNAs with sophisticated computer-aided design tools.
Figure 2. Schematics for framework nucleic acids. (a) The “framework” architecture of the Beijing National Stadium (Bird’s Nest). (b) FNAs are truly monodispersed nanostructures with precise molecular weights, which compare favorably with normal nanoparticles with more or less polydispersity. (c) Tailorable mechanical properties of FNAs by programming base composition and sequences.
caDNAno72 were developed. Both tools contain a graphical user interface and allow large-scale editing and visualization of strands and shapes. They can automatically generate staple sequences and extract the information for chemical synthesis, which minimizes tedious and error-prone tasks. Considering the large number of component strands in an origami object, in the two tools the visualization strategies were programmed to be flexible to allow the simplified display of a designed structure with key information and features. Perhaps caDNAno is more user-friendly, especially to the beginners, because of its easy-to-view and easy-to-operate three-panel interface (Figure 3a). In caDNAno, the left panel provides a cross section view (perpendicular to the helical axes) of the honeycomb or square helix lattice, with helices represented by circles; the middle panel allows in a side view for base-level editing of the connectivity of scaffold and staple paths, assigning sequences to scaffold paths, as well as exporting staple sequences; the right panel renders a real-time, 3D cylinder model for visualizing the constructed shape. Users get started by selecting helices in the left panel to approximate a 2D projection of the desired shape, and then move to the middle panel to work on the scaffold and staples. Following the online tutorials
3. FNA-ENABLED SPATIAL ORGANIZATION OF NANOSCALE OBJECTS Owing to the maturation of chemical synthesis of DNA, various chemical groups, including fluorescent dyes, thiols, biotins, etc., can be covalently linked with the component strands of FNAs at almost any position. And the precision assembly of FNAs can enable the precise spatial organization of molecules that are reactive to those chemical groups. The excellent chemical addressability, the precision assembly, and the high-ordered conformation together make FNAs perfect 17810
DOI: 10.1021/jacs.8b10529 J. Am. Chem. Soc. 2018, 140, 17808−17819
Perspective
Journal of the American Chemical Society
Figure 3. Representative bottom-up and top-down design procedures. (a) Bottom-up design: caDNAno interface (left) and exported schematic of example design (right). Adapted with permission from ref 72, Copyright 2009 Oxford University Press. (b) Top-down sequence design procedure for 3D DNA origamis of arbitrary shape. Adapted with permission from ref 73, Copyright 2016 American Association for the Advancement of Science.
squares, rectangles, triangles, five-point stars, etc., have been created with well-defined conformation in diameters of ∼100 nm, roughly 10-fold larger than the FNA objects created before.82 The staples, with an about 6 nm × 6 nm resolution, can be readily modified to conjugate biomolecules or nanoparticles for their precise docking on a 2D surface for functionalization: On the rectangle origami FNAs with appreciable surface areas, DNA walkers and nanomechanical devices were assembled to perform commanded and complex tasks in the nanoscale;83−86 and protein enzymes were organized with controlled interenzyme spacing and position to study the distance-dependent catalysis.87,88 The precision of organization and aptness of programming with origami FNAs also expedited the development of label-free detection of RNA89 and SNP41,81,90,91 (single nucleotide polymorphism). In addition, on the triangle origami FNAs, multiple gold and silver nanoparticles (AuNPs92 and AgNPs93) were discretely assembled with well-controlled orientation and
