In vivo production of RNA nanostructures via programmed folding of single-stranded RNAs

Mo Li; Mengxi Zheng; Siyu Wu; Cheng Tian; Di Liu; Yossi Weizmann; Wen Jiang; Guansong Wang; Chengde Mao

doi:10.1038/s41467-018-04652-4

In vivo production of RNA nanostructures via programmed folding of single-stranded RNAs

Li, Mo; Zheng, Mengxi; Wu, Siyu; Tian, Cheng; Liu, Di; Weizmann, Yossi; Jiang, Wen; Wang, Guansong; Mao, Chengde 2018-06-06 00:00:00 ARTICLE DOI: 10.1038/s41467-018-04652-4 OPEN In vivo production of RNA nanostructures via programmed folding of single-stranded RNAs 1 1 1 1 2 2 3 Mo Li , Mengxi Zheng , Siyu Wu , Cheng Tian , Di Liu , Yossi Weizmann , Wen Jiang , 4 1 Guansong Wang & Chengde Mao Programmed self-assembly of nucleic acids is a powerful approach for nano-constructions. The assembled nanostructures have been explored for various applications. However, nucleic acid assembly often requires chemical or in vitro enzymatical synthesis of DNA or RNA, which is not a cost-effective production method on a large scale. In addition, the difﬁculty of cellular delivery limits the in vivo applications. Herein we report a strategy that mimics protein production. Gene-encoded DNA duplexes are transcribed into single-stranded RNAs, which self-fold into well-deﬁned RNA nanostructures in the same way as polypeptide chains fold into proteins. The resulting nanostructure contains only one component RNA molecule. This approach allows both in vitro and in vivo production of RNA nanostructures. In vivo synthesized RNA strands can fold into designed nanostructures inside cells. This work not only suggests a way to synthesize RNA nanostructures on a large scale and at a low cost but also facilitates the in vivo applications. 1 2 Department of Chemistry, Purdue University, West Lafayette, IN 47907, USA. Department of Chemistry, University of Chicago, Chicago, IL 60637, USA. 3 4 Markey Center for Structural Biology and Department of Biological Sciences, Purdue University, West Lafayette, IN 47907, USA. The Institute of Respiratory Diseases, Xinqiao Hospital, 400037 Chongqing, China. These authors contributed equally: Mo Li, Mengxi Zheng. Correspondence and requests for materials should be addressed to G.W. (email: wanggs2003@hotmail.com) or to C.M. (email: mao@purdue.edu) NATURE COMMUNICATIONS (2018) 9:2196 DOI: 10.1038/s41467-018-04652-4 www.nature.com/naturecommunications 1 | | | 1234567890():,; ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04652-4 ucleic acids (DNA and RNA) have been extensively C) to a low temperature (e.g., 25 °C) over a long period of time. explored for molecular self-assembly and a wide range of Obviously, this thermal annealing process is not feasible for 1– Nnanostructures have been assembled from nucleic acids. nucleic acid self-folding in vivo. A potential approach is to design Such nanostructures can be applied to various ﬁelds from the targeted nucleic acid nanostructures both thermodynamically 10–13 physical devices to biomedical applications. Following DNA stable and kinetically favorable. To achieve this goal, the ssRNA is nanotechnology, programmed RNA self-assembly has rapidly designed to fold following a sequential and hierarchical pathway. evolved in hope that RNA has more structural complexity and Newly synthesized ssRNA would ﬁrst fold into hairpins while functional diversity. Up to now, DNA/RNA self-assembly gen- transcription. Hairpin structures are not only thermodynamically erally starts from chemically or enzymatically synthesized stable but also topologically simple. They only involve local single-stranded DNA or RNA (ssDNA or ssRNA). This method is interactions, thus, fold quickly. If any alternative structure forms, not desirable for large-scale production because of the excessive it would readily rearrange into the target hairpin structure via cost. A potential solution is to clone nucleic acids in bacteria, such local branch migration. Upon hairpin formation, which deﬁnes 14, 15 as E. coli, in the same way as recombinant proteins. How- the RNA’s secondary structure, most of the RNA residues are ever, DNA molecules normally exist as duplexes in the cell. inert as being in the content of duplexes, leaving a minimal Though direct cellular production of ssDNAs is possible amount of RNA residues as unpaired. The unpaired residues are 16, 17 in some special cases such as by reverse transcriptase or M13 able to further form long-range tertiary interactions, leading to bacterial genomes , there are limitations. For instance, the formation of fully folded, designed nanostructures. The the ssDNA length range is limited or additional enzymatic overall folding pathway is similar to that of the naturally occur- 18 21 treatment after puriﬁcation is needed. Comparatively, a more ring complex RNA structures, such as hairpin ribozymes. . suitable choice is RNA. Cellular RNAs primarily exist as single- Conceptually, the design concept resembles the principle that stranded, and their length can vary in a broad range. Recently, developed by Geary et al. However, a signiﬁcant change is that the Geary et al. have demonstrated that ssRNAs, in test-tubes, can short dovetail seams (2–3 bps) are avoided. Such short helical cotranscriptionally fold into designed nanomotifs, which can domains are not very stable and are likely to deformation under further assemble into 2D arrays. Delebecque et al. have cloned mild stress. RNA nanostructures in E. coli to organize chemical reactions in vivo. However, the nanoscaled, structural details of the RNA complexes have not been thoroughly characterized under native Results conditions. Molecular design. The RNA nanostructures in this study are In this work, we have developed a versatile strategy to prepare rationally designed based on natural RNA motifs and tertiary well-deﬁned nanostructures by folding individual long ssRNAs. interactions (Fig. 1), including: (i) RNA duplexes, (ii) RNA 22 23 Each nanostructure contains only one ssRNA molecule. The hairpins , (iii) 3-way junctions in open conformation (o3WJ) , resulting nanostructures can be cloned, expressed, and self-folded (vi) 3-way junctions in stacked conformation (s3WJ) observed in in E. coli. RNA nanostructures have been thoroughly character- the packaging RNA (pRNA) of phi29 bacteriophage, (v) ized by gel electrophoresis, atomic force microscopy (AFM) coaxially stacked kissing loops (KLs) found in the dimerization imaging, and cryogenic electron microscopy (cryoEM). initiation sites of HIV-1 RNA, (vi) a 3-way loop (3WL) inter- A key challenge of this approach is to design the folding action observed in phi29 pRNA, (vii) 4-way junctions in open pathway to avoid kinetic traps. For nucleic acid self-assembly, the conformation (o4WJ), and (viii) 90°–kink found in the internal target structures are designed to be thermodynamically stable, but ribosome entry site (IRES) of the hepatitis C virus (HCV) RNA often not kinetically favored. This problem is commonly solved genome. According to the molecular designs, DNA templates by slowly cooling the samples from a high temperature (e.g., 95 ° coding for the ssRNAs have been synthesized so that RNAs can Duplex Hairpin 90°-kink Kissing-loops (KL) Open 3-way junction (o3WJ) Open 4-way junction (o4WJ) Stacked 3-way junction (s3WJ) 3-way loops (3WL) Fig. 1 Component motifs of RNA structures. For each motif, a schematic drawing and a 3D model are shown. The thick colored lines and thin gray lines represent RNA backbones and basepairs, respectively 2 NATURE COMMUNICATIONS (2018) 9:2196 DOI: 10.1038/s41467-018-04652-4 www.nature.com/naturecommunications | | | ~10 nm ~10 nm NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04652-4 ARTICLE be produced by in vitro transcription. Alternatively, the DNA RNA square and double-square.We ﬁrst tested this approach by templates can be inserted into a plasmid and introduced into E. the assembly of two 2D structures: a square (S) and a double- coli. Then cellular transcription machinery will transcribe the square (S2). Figure 2 shows the designs and experimental char- gene into the corresponding ssRNA, which spontaneously folds acterizations (For RNA sequences and 3D models, see Supple- into the designed nanostructures inside the cells. mentary Figs 1 & 2). RNA S contains four 90°-kinks and a KL 3′ 2+ Mg 5′ 3′ 2+ Mg 5′ S2 ~17.5 nm cd S* S* 300 bp 220 bp 100 bp S2 300 bp 220 bp S2 100 bp Fig. 2 Designs, folding, and characterization of an RNA square (S) and an RNA double square (S2). a, b The molecular design and single-stranded folding pathways for S and S2, respectively. The RNA single strands are colored in a rainbow gradient from 5′ (red) to 3′ end (purple). Red, green, and blue boxes with dashed lines highlight a 90°-kink, a KL interaction, and a 3WL interaction, respectively. Each edge is composed of a two-turn RNA duplex. Characterization of c, d RNA square and e, f RNA double square. c, e electrophoretic analysis; d, f Atomic force microscopy (AFM) imaging. Note that S* has the same sequence as S except that one loop sequence is altered so that no KL interaction can form. (Scale bar: 20 nm) NATURE COMMUNICATIONS (2018) 9:2196 DOI: 10.1038/s41467-018-04652-4 www.nature.com/naturecommunications 3 | | | DNA markers DNA markers Quenched Quenched S* Annealed S Annealed As transcribed S* As S* transcribed ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04652-4 interacting pair, while S2 contains four 90°-kinks and two 3WL potential of expressing RNA nanostructures inside cells. It is pairs. After in vitro transcription, puriﬁed RNA molecules were worthy noting that for single-component RNA folding method, thermally annealed (slowly cooled from 65 °C to 4 °C over 2 h in a any experimental issue related to stoichiometry is avoided. 2+ neutral, Mg -containing, aqueous buffer) to promote the RNA Direct visualization by AFM conﬁrms that the RNA molecules 2+ folding. RNA molecules ﬁrst folded into Mg –independent, indeed fold into the designed structures (Fig. 2d, f). In the AFM intermediate, loose secondary structures, which then further fol- images, randomly distributed RNA particles exhibited uniform 2+ ded into Mg –dependent, compact, tertiary structures via loop- sizes. Their apparent shapes are closed and consistent with the loop (KL or 3WL) interactions. The resulting RNA square (S) designs: S particles as squares and S2 particles as double squares. samples were analyzed by native PAGE (Fig. 2c). A dominant In contrast, the control molecule, S*, appears as a very different, sharp band is observed; suggesting that the RNA square (S)is open geometry because the KL interaction is disrupted. Please fully folded. To show the critical role of the KL interaction in the note that AFM images give a direct visualization of the RNA folding, a control molecule (S*) is prepared, in which one loop structures. However, discrete frameworks could be easily sequence is altered so that the KL interaction is interrupted disrupted during sample preparation and by AFM probes, (Supplementary Fig. 1). S* migrates slightly slower than S, indi- resulting in artiﬁcial structural heterogeneity. Thus, the RNA cating that S* can fold into the open intermediate secondary folding efﬁciency should not be estimated by AFM imaging of a structure, but not the compact tertiary structure. The formation small number of objects; instead, native PAGE of the bulk sample of S is a unimolecular process and is expected to be kinetically gives a reliable estimation. fast. To prove this hypothesis, we performed a quenching experiment by plunging the RNA solutions from 65 °C onto the RNA nanoﬂower and tetra-square. To demonstrate that this ice. In the native PAGE, the quenched samples (both S and S*) approach is versatile and could be used to assemble large, com- migrate identically to those annealed samples, indicating that the plex geometry, a 5-petal nanoﬂower and a tetra-square (S4) were intramolecular folding of the designed RNA molecules is indeed a designed (Fig. 3, Supplementary Figs 3 and 4). The RNA ﬂower is fast process. In addition, the comparison between S and S* 1571 bases long and the ﬁnal structure contains ﬁve petals. The molecules gives an estimation of the folding yield of S close to folding pathway is illustrated in Supplementary Fig. 3b. The RNA 100% as no RNA in the lane of S migrates like the partially folded quickly folded into an intermediate, highly branched, secondary S*. More strikingly, the RNA molecules could spontaneously fold structure, which is composed of multiple duplex regions and into the designed structures cotranscriptionally. After transcrip- multiple loops. Via KL interactions, the branch ends pairwisely tion, the crude RNA samples were directly analyzed by PAGE associated with each other to form the ﬂower petals. Note that ten without any puriﬁcation or thermal treatment (the lane as indi- 90°-kink loops, six KL interactions, and ten 3WLs taken from cated As Transcribed). The migration pattern was identical to phi29-pRNA were integrated into the RNA strand to facilitate the those of puriﬁed and thermally treated samples, indicating that, as desired folding. The expected ﬁve-petal-ﬂower particles were expected, the RNAs folded into the designed structures even in clearly observed under AFM (Fig. 3b), conﬁrming that the RNA crude solutions without any annealing step. A similar result was indeed folded into the designed structure. In the transcription also observed for S2 (Fig. 2e). This result points to a great mixture from a crude PCR mixture (no puriﬁcation for DNA ab Transcription mixture Purified RNA cd Transcription mixture Purified RNA Fig. 3 Folding of complex RNA nanostructures from single RNA strands. a Structural design of an RNA 5-petal ﬂower, which contains six KL interactions, ten 90°-kinks, and ten 3WJs. b AFM images of the RNA ﬂowers. Left: transcription mixture from a PCR mixture. Right: thermal annealed, puriﬁed RNA nanoﬂower molecule. Scale of inset: 60 nm. c Structural design of an RNA tetra-square (S4), which contains four 3WLs, four 90°-kinks, and a 4WJ. d AFM images of the RNA S4. Left: transcription mixture. Right: thermal annealed, puriﬁed RNA S4. Scale of inset: 25 nm. (Scale bar: 50 nm) 4 NATURE COMMUNICATIONS (2018) 9:2196 DOI: 10.1038/s41467-018-04652-4 www.nature.com/naturecommunications | | | NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04652-4 ARTICLE Fig. 4 Folding of a single-stranded, 4-turn, RNA tetrahedron (T4). a Structural design of T4. An RNA single strand is rainbow colored from 5′ to 3′ end. It contains four 3WJs and will fold from a 2D branching structure 5′ into a 3D tetrahedron upon the three pairs of KL interactions (indicated by 3′ KL T4 dashed, double-arrowed lines). Each edge is four helical turns long. Scale Interaction bar, 50 nm. b A control molecule T4*. One loop sequence is altered to prevent one KL interaction. Thus T4* will assemble into a ﬂat, double- triangular shape instead of a tetrahedron. The models were built with 39 37 Coot and Chimera . All KLs are colored orange in the model. Scale bar, 50 nm. c, d AFM images of T4 and T4*, respectively. For each structure, three particles are zoomed-in and ﬁtted with corresponding shapes. e, g 5′ Cryogenic electron microscopy (cryoEM) characterization of T4. e A raw 3′ KL T4* cryoEM image. Each white box indicates an individual RNA particle. Scale Interaction bar, 50 nm. f Four different views of the reconstructed structural model of the RNA tetrahedron (top) and corresponding views of the simulated model (bottom). Scale bar, 5 nm. g Pairwise comparison between raw cryoEM images of individual particles (left) and the corresponding projections (right) of the reconstructed structural model. The raw particles were T4 selected from different images to represent views at different orientations fold with high yields. And this strategy was further conﬁrmed by the success of the folding of an RNA tetra-square (S4) structure (Fig. 3c, d and Supplementary Fig. 4). RNA tetrahedron. One important test of molecular self-assembly is to generate discrete, 3D nanostructures, which can be readily T4* achieved by the reported strategy (Fig. 4). For demonstration, a tetrahedral structure (T4) is designed to fold from a 623-base- 2 long ssRNA. It contains six edges and four vertices. All edges are four helical turns long. Three of them are standard A-form duplexes and each of the other three contains a KL interaction in the middle. Each of the vertexes is an o3WJ containing four unpaired uracils on each strand at the center to ensure sufﬁcient out-of-plane ﬂexibility to fold. Upon cooling, the ssRNA ﬁrst folded into a three-branched structure (Fig. 4a, left), and then closed into a tetrahedral geometry via KL interactions (Fig. 4a, right). To show the importance of the KL interactions, a control molecule (T4*) was prepared, in which a pair of KL interaction was disrupted by sequence mutation. It is expected to only fold into an open, 2D, double-triangular shape instead of the compact, fully folded, 3D tetrahedron (Fig. 4b). We have characterized the tetrahedron by native PAGE (Supplementary Fig. 6), AFM imaging (Fig. 4c, d), and cryoEM (Fig. 4e–g). PAGE analysis suggests that RNA molecules efﬁciently fold into the designed structures (Supplementary Fig. 6a). Each sample appears as a sharp, dominant band in the corresponding lane. The migration pattern is consistent with the expected structures. The fully-folded, compact, 3D tetrahedron (T4) has a higher mobility than the open, 2D, double-triangle structure (T4*). The folding is fast, and no difference has been observed between quenching and annealing for each sample. In the lane of T4, no RNA migrates like the partially folded T4* molecule; suggesting that all T4 molecules are fully folded. In order to conﬁrm the migration pattern and the folding ability, we have prepared two more RNA tetrahedrons with different sizes templates and RNA), some linear structures (150–400 nm long) and their corresponding control molecules: a three-turn-edged were observed. They were unfolded or partially folded RNA tetrahedron (T3 and T3*, 484 bases long, Supplementary Fig. 5) molecules, or truncated RNA molecules and DNA templates (the and a two-turn-edged tetrahedron (T2 and T2*, 340 bases long, designed DNA template is 1591 bps or 540 nm long). In the Supplementary Fig. 5). Similar migration patterns are observed puriﬁed RNA samples, only the 5-petal nanoﬂowers remained in (Supplementary Fig. 6b-c). AFM imaging gives more direct the AFM image. Such a complicated pattern would be quite structural clues. T4 samples appear compact and consistent with challenging for the traditional tile-based method, which would a collapsed tetrahedral geometry (Fig. 4c), while T4* samples involve multiple different tiles and sophisticated inter-tile inter- exhibit clear double-triangle structures (Fig. 4d and Supplemen- actions, generally leading to a low assembly yield. With this tary Fig. 7). Please note that some T4 structures are disturbed ssRNA folding strategy, complicated structures could quickly self- during AFM imaging. NATURE COMMUNICATIONS (2018) 9:2196 DOI: 10.1038/s41467-018-04652-4 www.nature.com/naturecommunications 5 | | | ~11 nm ~11 nm ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04652-4 CryoEM study revealed the native, 3D, tetrahedral structure of IPTG), abundant double-square shaped particles are clearly T4. In raw images of the T4 sample (Fig. 4e), uniform-sized identiﬁed, which are identical to the S2 sample prepared from particles could be easily identiﬁed (indicated with white boxes). in vitro transcribed RNA (Fig. 5c: samples from phenol lysis Their shapes resembled the 2D projections of a tetrahedron. method; Supplementary Fig. 9d: samples from sonication From the observed individual particles, a technique of single- method). In contrast, no double-square shaped particles are particle 3D reconstruction was applied and resulted in a observed from the cells without the S2 gene or the cells without structural model of a tetrahedron at a resolution of 4.1 nm expression induction. Based on the band intensities in the PAGE (Fig. 4f). To verify the reconstructed model, two comparisons graph (Supplementary Fig. 9b), we can estimate that S2 molecule were performed: the computed 2D projections with individual accounts for roughly 11.2% of all RNAs in the cell; suggesting that raw particles (Fig. 4g) and with class averages of raw particles in vivo expression is potentially an efﬁcient way to produce RNA with similar views (see Supplementary Fig. 8). Clear similarities in nanostructures. Similarly, we can in vivo express the tetra-square the pairwise comparisons veriﬁed the reconstructed model. Note (S4) structure (Supplementary Fig. 10). It seems that the RNA that the six struts in the model were not the same. folding is efﬁcient, but the RNA transcription yield is low or the RNA is not stable in the cell. In the future, it might be worthy using natural RNAs, e.g., tRNA or 5 S rRNA, as scaffolds to In vivo production of RNA nanostructures. The most promis- improve the RNA production yields of longer and complex ing feature of the current approach is that it is compatible with 14, 15 strands. in vivo expression of the designed RNA nanostructures. In our design, each RNA nanostructure contains one single strand. Therefore, an RNA strand, in the same way as peptides fold into Discussion proteins, can spontaneously fold into the designed structure In vivo production of nucleic acid (DNA/RNA) nanostructures is under the physiological condition during in vivo transcription. To a major challenging to the ﬁeld of nucleic acid nanotechnology demonstrate this capability, we have cloned and expressed the and has attracted signiﬁcant efforts. The original idea was ﬁrst double square structure (S2)in E. coli (Fig. 5). Brieﬂy, the S2- proposed by Nadrian Seeman in 1997. But the complicated coding DNA sequence is inserted into a pET23a plasmid, a topology associated with that design prevented it from experi- bacterial expression vector, under the control of a T7 promoter mental realization. The ﬁrst experimental progress was elegantly (Supplementary Fig. 9a). The resulting recombinant plasmid is done by Shih et al. in 2004. They assembled an octahedron from then transformed into BL21 (DE3) E. coli cells and the RNA a long enzymatically generated DNA single strand and a few transcription is induced by isopropyl β-D-1-thiogalactopyranoside helper strands. In 2006, Paul Rothemund developed a powerful (IPTG). Then the cells were lysed by phenol or sonication. and generally applicable strategy, so-called DNA origami. It fol- Without any further manipulation, the aqueous solution of the ded long, biologically produced, M13 bacteriophage genome by total cell lysates was directly examined by both PAGE and AFM hundreds of chemically synthesized, short DNA strands into imaging. On native PAGE, the desired RNA molecule from E. coli designed nanostructures. In 2008, Lin et al. cleverly used bacteria migrates the same as the in vitro prepared S2 sample (Supple- to produce simple DNA nanostructures, such as a DNA four way mentary Fig. 9c), suggesting that the RNA from the cells is well junction. In 2011, Delebecque et al. produced RNA molecules, folded as the expected double-squared structure. This is further which further associated with each other to organize chemical conﬁrmed by AFM imaging (Fig. 5c and Supplementary Fig. 9d). reactions in E. coli. In 2011, Afonin et al. reported in vitro, In the raw cell lysate after expression induction (addition of co-transcriptional assembly of multi-stranded RNA complexes. a b ~10 nm E. coli – Plasmid – IPTG + IPTG Fig. 5 Cloning and in vivo expression of RNA nanostructures. a Overall scheme. An expression vector (pET23a, orange) carrying an S2-coding DNA sequence (green) is transformed into E. coli cells. Upon IPTG induction, the S2 gene gets transcribed into RNAs, which self-fold into the designed double- square structure in the cell. b Structural model of the S2 structure. c AFM imaging of cell lysates from E. coli without the expression vector (- Plasmid), or with the vector but without S2 gene expression induction (- IPTG), or with S2 gene expression ( + IPTG). (Scale bar: 20 nm) 6 NATURE COMMUNICATIONS (2018) 9:2196 DOI: 10.1038/s41467-018-04652-4 www.nature.com/naturecommunications | | | ~17.5 nm NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04652-4 ARTICLE Table 1 Sequence of DNA templates S5′-TTCTAATACGACTCACTATACCGTGTACTCGGAGGAACTACTCATAT TCGAGCGTCGGAGGAACTACTGCAGTGGCTCGTTATTCGG AACTACTATGGTGAAGGAGGCACGCCATAGCCGAATAACGAGCCATTGCAGCCTCCGACGCTCGAATGTGAGCCTCCGAGTACACGGGCTG AGCCACGTGAAGCCTCCACGCGTGGAACTACTCAGCAGTGCTAAGAATCTA-3′ S* Identical to S except the underlined sequence was replaced by AAAAAAAAA S2 5′-TTCTAATACGACTCACTATAGACTCGGAGGAACTACTCACATTC GAGCGTTGATGGAGGTCAACGCTCGAATGTGAGCCTCTGAGTCCT GTCGATCAGCCTGGTCTAGCATCTATCAGCCCAGATTCAACTATTGGGGATACTCCCTGAATCGGATCACGCACGGAACTACTGTAGTGCTT CGTTAGTCGGAACTACTGCTCTAGATAGGTTGATGTCGGTCAACCTATCTAGAGCAGCCGACTGACGAAGCATTACAGCCGTGTGTGATCCG ATTCAGCCCAGTAGTTGAATCTGGGGATACGACCTGATAGGTGCTAGGCCAGGAACTACTGATTGACAGGCAGTGCTAAGAATCTA-3′ FL 5′-TTCTAATACGACTCACTATACTCTAGAAGACCGAGGCTGGCTATAT GAGGAACTACTCGGTGAAGAGGAGACGCCGAGCCTCATGTAGCCAGTCTCGGTCTCCGTCGTCAGACACTCGGGTGTGTGTCGTTGCGTCG ACGGGTTGTGTGTCCTGTCTTGGCAGGTCTGGACCGGGAGCGCTTGAGCTCAGCCATTTGAACTCCTCACGAATGGAACTACTGAGTTCAA GTGCTCCTGGTCCGGACCTGTCAAGACGGGACACTTTGTAGGGTGGTGTTTCACGGGACATCACCTGTGGAGGGTTATCGGTACTCTGGG CACGGAACTACTGCATGAATCGAGCACGTGCAGCCGTGCTCAGAGTGCCGATGACCCTCTACAGGTGGTGTCCCGTGTGTAACGTCTCGAA GAGGAGGTGTGTCAAGTAGGCACAGCTGAGGTTCGAACGGACTTCGGCTCCCAGCCTGGTGAAGCTCGAACGCCAGGAACTACTGGGAG TCGAAGTCTGTTCGAATCTCAGTTGTGCCTGCTTGACTTTGTCAGGGAGTGTTTCAGTAGTAACGGGCCGACTAACTGGAGCGTGGGAAGA TCATTGGAACTACTAGCTGAAACGGCAACGGCTAGCCAATGGTCTTCCTACGCTCTAGTTGGTCGGTCCGTTGCTACTGTGTGTTTGAACAC TGGACGAACACTACTCCCTGATCACCTTCTCTTTGAGATGTTACTACCACCCTATCAACCTGTCGGCGCAATGACACTACCCTGGGATCAGT CTGATCTATCGTGACGACTAGAGCTCGATCAACAGGCTATTCGGTCAAGACTACTCTCAGATCACGGTGAACCAGTGACGCCGTGTGTCGA GTCAACTGCGTTGCGTCCGTATCACTACCTCTTGGAGAGCCCAGTGAATGCCGTACGCTGGGAACTACTCTCCGAGAGGTGGTGATGCGGA TGCAACGTAGTTGGCTCGACTTTGTCTGAGAGTGTTTCAGCATGCGTCAGCTCTGTGGTGAATTCAGCTAGCACTGTCAGGAACTACTCAGT GAAGATAGGACGCTGAGCCTGACGGTGCTAGTTGAATTCATCACAGGGCTGATGCATGCTGTGTCTTGGCCGAGTAGCTTGTTGTGTCGGT GTCACCGTATGGACATCTGTCTTCAGCATTGGCCCTAGCCGGCTGAACCTATCACGGCCGGAACTACTAGGGTCAATGCTGGAGACAGGTG TCCATGCGGTGGCACCGACTTTGTCGAGCTCTGTTTCACTAGGAGGATCCGAGCATTCTTCACTGGGGTTCTCGTGGTGGAACTACTCACCG AAGCACGTACGGTGAGCCACCATGAGAACTCCAGTGGAGAATGTTCGGATTCTCCTAGTGTGTCGTCGCGATGGATCGGACTGTGTGCCTG GACGCAGTCCCTTACGCTCTGCTAGGGTGCTACTTAGCCAAGTGAAACGTGCACGCTTGGAACTACTAAGTGGCACCTTAGCGGAGCGTGA GGGATTGCGTTCAGGCACTTTGTCCCAGGGTGTTTCACCCGGGTGTCTGATGACGGACTAGCATAACCCCTTGGGGCCTCTAAACGGG-3′ S4 5′-TTCTAATACGACTCACTATAGGTCTAGAGGTGAGTGTGTAACTACG GTCTAGTCGTGGAGGAACTGGTCGTAGTTCCTCCGCGACTAGGCCGACACACTCACCCATCACTCCAGCCACTTCCAGGGAAACTCCGGCTT CGTGTGATCTCGACATTCTGGGGATACGACCTGGTTACCCTCCTAGACTATCAACTAGGTGGGAGTGAATGCTGTTCTGGGAGTAGAACAGC GTTCACTCTCACCGATAGTCTGGGAGGGTAATCAGCCCAGAGTGTCGAGATTACACGTTCGTACAGACACGTCTGAGTGGGGATACTCCCT GAACTCAGGTGCTTGAGCTCAACTAGGAGTTGAGTCTTCTCAACGTGTTGCTACGTTGAGAAGATTCAACTCCGAGTTCAAGCACCTGGGTT CAGCCCACTCAGGCGTGTCTGTGCGTTGTGAGTATCGACAGGATAGAGGGATAGCAACGTGCAGTGGAGAACATTACACAACTAGTGTCAG AGCTAGCAGGTCGTGAGCGTACGACCTGCTGGCTCTGACACGTGTGATGTTCTCCATTGCACGCCTCTGTCCTGTCGATGCTCACTTGCTG GAGTTTCTCTGGAAGTGGGATACGCTCTGGAGTGATGCCTTGGGGCCTCTAAACGGG-3′ T4 5′-TTCTAATACGACTCACTATAGACGTGTAGTTGTCTGTTACCTTTTTA GCGTGTAGATAGTCGCTACAAAAGCGGTATGTAGCGACTATCTACGCGCTTTTTAATCAGACGCAAGTTTATCAACACCTGAGATAGACTTG CGTTTGATTTTTTAGGTAACAGACAACTACATGTCCCCAATCTAGTCTCAGTCCGGATTTTGAGGATGAGCACATTTGCGCGTATGTGGTCC GAAGGTCATGGCATTTTTCCGTACATCAACTTAGGTGCAACAGGTGAGCATCTAAGTTGATGTACGGATTTTGTGCGTCGCTCGACGGCCGAA ACGGCAACGGTCGTCGAGCGATGCACTTTTTGCCATGACCTTCGGACCACATACGCGCGAATGTGCTTATCCTCTTTTTCATGCGATGGCT CGGCGGCTCTTGGTACATTCTGGCTCGACTTTTTTAGGCTTCTCTGTCACGGATGCAATGCCGTAGCATTCGTGACAGAGAAGCCTTTTTAG GTTAACTACACGGATATAAACCGCTAATATCTGTGTAGTTAGCCTTTTTAAGTCGAGCCAGAATGTACCAAGAGCCGTCGAGCCATTGCATG ATTTTTCCGGATTGAGACTGGATTGGGGCAGTGCTAAGAATCTA -3′ T4* Identical to T4 except the underlined sequence was replaced by AATTTTTTA T3 5′-TTCTAATACGACTCACTATAGACGTGTAGTCGCCTGTTACCTTTTTA GCGGCTACAAAAGCGGTATGTAGCCGCTTTTTAATCAGACGCAAGTTTATCAACACCTGAGATAAACTTGCGTCTGATTTTTTAGGTAACA GGCGACTACACGTCCCCAATCCGGATTTTGAGGATGAGCACATTCGCGCGTAGGTCATGGCATTTTTCCGTACTGCAACAGGTGAGCAGTAC GGATTTTGTGCATCGCTCGACGGCCGAAACGGCAACGGCCGTCGAGCGATGCACTTTTTGCCATGACCTACGCGCGAATGTGCTCATCCT CTTTTTCATGCAATGGCTCGGTAATTCTGGCTCGACTTTTTTAGGCTAATGCAATGCCGTAGCATTAGCCTTTTTAGGCTAACTACACGGAT ATAAACCGCTAATATCCGTGTAGTTAGCCTTTTTAAGTCGAGCCAGAATTACCGAGCCATTGCGTGATTTTTCTGGGTTGGG-3′ T3* Identical to T3 except the underlined sequence was replaced by AATTTTTTA T2 5′-TTCTAATACGACTCACTATAGGCTACTGTTTATAGCGAGGCCGTG ATTCGCTTTTTAAAAGTCATAATGCCGTAATGACTTTTTTTTTCGA GCCGAATCGCCAACGGCTCGAATTTAAGCGAATCACGGCCTCGCTATTTTAATCGATGCATCTGACTGCTCTTTTAGACGCUCGAATGGCGA ACGGGCGTCTTTTAGTACGTCCAAGTCCACAGGACGTACTTTTAGAGCAGTCAGATGCATCGATTTTTCAGTAGCCGCATTGUATACTGCTT TCTGAGTACCAAGTGGACAGGTACTTAGTTTCGGTGCGGCAAACGGCAAGCCGCACCGTTTGCAGTATGCAGTGCTAAGAATCTA-3′ T2* Identical to T2 except the underlined sequence was replaced by AATTTTTTA S2’ 5′-TTCTAATACGACTCACTATACTCTAGAACAGGCGGAGGAACTACTC ACGTTCCAGCGTTCATGGAGGTGAACGCTGGAATGTGAGCCTCTGCCTGCTCCTGATCAGCCACGTGTGACATCTAAGAGCCCATCCTCACC AGTAGGGGATACTCCCTGAATCGGAACATGGACGGAACTACTGTAGTGCTTGGTTAGTCGGAACTACTGCACGAGGTAGGTTGATGTCGGTC AACCTATCTCGTGCAGCCGACTGACCAAGCATTACAGCCGTCCGTGTTCCGATTCAGCCCTGCTGGTGAGGATGGGGATACGACCTCTTAG GTGTCACGCGTGGAACTACTGATTAGGAGCTAGCATAACCCCTTGGGGCCTCTAAACGGG-3′ T7 promoter sequences are highlighted in bold Two restriction enzyme sites, XbaI and StyI are formatted in italics (for in vivo double-square S2′ and for tetra-square S4) (FL: RNA nanoﬂower, S2′: double-square for in vivo expression) In 2013, Ducani et al. enzymatically, either in vitro or in vivo, in vitro fold long RNA strands into nanostructures. This strategy produced ssDNA, which were released with restriction enzymes is closely related to our work. In 2016, Elbaz et al. took advantage in test-tubes and used to assemble DNA nanostructures with of reverse transcriptase to in vivo produce ssDNAs, which could 30 31 thermal annealing. In 2014, Geary et al. developed a strategy to in vivo assemble into multi-stranded DNA nanostructures. In NATURE COMMUNICATIONS (2018) 9:2196 DOI: 10.1038/s41467-018-04652-4 www.nature.com/naturecommunications 7 | | | ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04652-4 2017, Hendrik and coworkers used bacteria to cost-effectively and Native PAGE. 4% native PAGE gel was prepared with 19:1 acrylamide//bisacry- 2+ lamide gel and TAE/Mg buffer. The gel was run at 4 °C. Then stained with at large-scale produce single DNA strands, which in vitro self- Stains-All (Sigma) and scanned by an HP scanner (Scanjet 4070 Photosmart). assembled into multi-stranded, DNA nanostructures upon ther- mal annealing. In 2017, Yan and coworkers enzymatically Agarose gel analysis. Samples were loaded into 1% agarose gel and run in TBE produced DNA or RNA single strands, which, upon thermal buffer on a FB-SB-710 electrophoresis unit (FisherBiotech) at room temperature annealing, in vitro folded into designed nanostructures without under constant voltage of 80 V. The gel was stained in EB and illuminated under topological tangling. UV light. Images were taken by cell phone. In conclusion, we have developed a programmable ssRNA folding strategy for the preparation of complex nanostructures. AFM imaging. Mica surface (Ted Pella, Inc.) was pre-treated with poly-L-lysine (0.1% w/v, Ted Pella, Inc.) to increase the adsorption of small RNA particles. 20 µL Each nanostructure is primarily composed of RNA duplexes and poly-L-lysine (5 µg/ml) was added onto freshly cleaved mica substrate and was is folded from one long ssRNA. Thus, the formation of the incubated for 30 s. Then the surface was washed by 100 µL H O and dried by nanostructures is independent of stoichiometry and RNA con- compressed air. Annealed RNA solution was diluted to 50 nM and a drop of 5 µL 2+ centration. This strategy allows cloning and expression of RNA was deposited onto pre-treated mica and incubated for 2 min. 20 µL TAE/Mg nanostructures inside of cells in the same way as those of buffer was further added. Imaging was performed in a ﬂuid cell under tapping mode on a Multimode AFM (MMAFM 2, Veeco) with SNL-10 probe (Veeco, Inc.). recombinant proteins; pointing to a cost-effective way for large- scale production of nucleic acid nanostructures. We expect this Cryo-EM imaging.3 μL RNA solution (50 nM) was deposited on the Lacey strategy would facilitate a wide range of in vivo biomedical Carbon-only grid (300 mesh Cu, Ted Pella, Inc.). The grid was blotted with two applications of RNA nanotechnology, such as integrating multiple pieces of ﬁlter paper (Whatman, 1001-055) and was immediately ﬂash frozen in functional RNA moieties to regulate cellular processes. liquid nitrogen-cooled liquid ethane. Images were taken on FEI Titan Krios with − 2 accelerating voltage of 300 kV under 90 e /Ǻ dose condition at under-defocus 8 μm to enhance image contrast. Methods Oligonucleotides. Double-stranded DNA templates were purchased from IDT, Inc. Polymerase chain reaction was applied to amplify DNA duplexes into large Single particle reconstruction. The single particle reconstruction was carried out amount. The sequence of DNA templates can be found in Table 1. using computer software EMAN 2. Totally, 470 particles were selected to go Sequence of PCR primers: through reconstruction process using EMAN2 reconstruction package, following P1: 5′-TTCTAATACGACTCACTATA-3′ the standard protocol. (Wiki at: http://blake.bcm.edu). A general summary of the P2: 5′-TAGATTCTTAGCACTGC-3′ steps includes: (1) Import and convert CCD micrographs; (2) Pick typical particles; P3: 5′-CCCGTTTAGAGGCC-3′ (3) CTF determination; (4) Remove bad particles and build particle sets; (5) Make P4: 5′-CGTTATTGTACCCAACCCAGAAAAATCAC-3′ 2D reference-free class-averages; (6) Build an initial 3D model; (7) 3D high reso- S/S1*/S2/T2/T2*/T4/T4*: P1 + P2; FL/S2’/S4: P1 + P3; T3/T3*: P1 + P4. lution reﬁnement; (8) Evaluate the resolution of 3D map. About 470 randomly selected particles were used to build 12 class averages, which were used to generate initial models under different symmetries, such as C3 symmetry, TET symmetry or DNA sequence design. When designing RNA sequences, we take two parts into without symmetry (see Supplementary Fig. 8). Totally around 360 particles were consideration. (i) The single stranded loops in the RNA secondary structures, used for 3D reﬁnements of the RNA tetrahedron. The reﬁnement was carried out including the 90°-kink loop, KLs, pRNA interacting loops. They are critical regions with a 2° angle interval under C3 symmetry. A projection matching algorithm was and determine the long-range interactions and local conformations. For the 5-nt- applied for the determination of the center and orientation of raw particles in long, 90°-kink loop, the sequence is strictly taken from the HCV IRES domain II. iterative reﬁnement. After reﬁnement, the class-averages of raw particles and the For KLs and pRNA loops, the exact sequence is not important, but sequence 2D projections of 3D model were generated in two ﬁles but paired in order. The complementarities are ensured for the interacting pairs. (ii) The remaining duplex comparisons shown in the result were manually selected to present the distinct and regions. They provide all the struts in the structures and do not involve long-range comprehensive orientations. The resolution of resulted structural density map was interactions. Their sequences are randomly generated by a computer software determined to be at 4.1 nm. Final reconstruction results were visualized by UCSF Tiamat. Roughly 10% of the Watson-Crick basepairs (A–T and G–C) are replaced 37 Chimera software . by G–U pairs to facilitate PCR ampliﬁcation of the DNA templates. Such G–U pairs are evenly distributed along the duplex regions. Finally, the RNA sequences are checked by a computer software Mfold to make sure that the RNA will fold into Recombinant plasmid preparation. Double-stranded DNA template and pET23a plasmid (GenScript Inc.) were double-digested by XbaI and StyI-HF restriction the designed secondary structures. If there are any undesired, strong secondary structures, the sequence will be manually altered to avoid such structures. enzymes (New England BioLabs Inc.) by following the manufacturer- recommended protocol. The reaction was performed in 100 µL CutSmart buffer at 37 °C for 2 h, and 20 units of each enzyme were added. After digestion, DNA was Polymerase chain reaction. DNA double-stranded template was ampliﬁed by recovered by DNA clean & concentrator columns (Zymo Research Corp.) to get rid polymerase chain reaction using Taq DNA polymerase kit (New England Biolabs of proteins. Then the template and plasmid were ligated by T4 DNA ligase (New Inc.). 4–10 ng DNA template, 200 µM dNTPs, 200 nM of forward and reverse England BioLabs Inc.), and were incubated overnight at 16 °C. primers were dissolved in 100 µL standard Taq reaction buffer. Then 0.5 µL Taq DNA polymerase was added to the system. The solution was initially denatured for Plasmid cloning and bacteria culture. Ligation product was combined with XL1- 30 s at 95 °C, and started 30 cycles at 95 °C for 30 s, 50 °C for 1 min, and 68 °C for Blue competent Escherichia coli (E. coli) cells for transformation. The mixture was 2–3 min. Then kept the ﬁnal extension at 68 °C for another 5 min. PCR product kept on ice for 30 min, and then heat to 42 °C for 90 s and put on ice for 2 min. was stored at 4 °C for short time storage and at −20 °C for long terms. Spread the transformed bacteria on Lysogeny Broth (LB, 100 µg/ml ampicillin was added) (RPI. Corp.) plates and incubate overnight at 37 °C. After the colonies In vitro RNA preparation. RNA molecules were synthesized from in vitro tran- growing into desired size, inoculate a colony into an individual aliquot of 6 ml LB/ scription using T7 RNA polymerase (AmpliScribe T7-Flash transcription kit; ampicillin liquid medium and shake overnight at 37 °C. 1.5 ml bacteria solution was Epicenter, Inc.). Corresponding DNA template was added, and the experiment was taken out into another 6 ml fresh LB/ampicillin liquid medium and kept shaking at conducted by following the manufacturer-recommended protocol. 37 °C until UV absorbance was around 0.5 OD at 600 nm, measuring by a UV/Vis spectrophotometer (Beckman coulter, DU520). Plasmid was extracted using plas- mid isolation kit (Zymo Research Corp.). The puriﬁed plasmid was characterized by 2+ Formation of ssRNA structures. ssRNA was prepared in TAE/Mg buffer and restriction enzyme digestion and gel analysis to conﬁrm the correct insert size. annealed by different processes: Quenched---65 °C/5 min, 0 °C/5 min; Regular anneal---65 °C/5 min, 50 °C/30 min, 37 °C/30 min, 22 °C/30 min, and 4 °C/30 min. 2+ TAE/Mg buffer contained 40 mM Tris base (pH 8.0), 20 mM acetic acid, 2 mM In vivo RNA expression induction. The recombinant plasmid was transformed TM EDTA, and 12.5 mM magnesium acetate. into BL21 Star (DE3) competent E.coli cells (Life Technologies Corp.) for RNA expression, following the manufacturer’s protocol. The cell culture process was the same as described above. When UV absorbance was around 0.5 OD at 600 nm, Denaturing PAGE. 5% denaturing PAGE gel was prepared with the 19:1 acryla- IPTG (1 mM) was added for induction and the solution was shaken at mide/bisacrylamide solution, 8 M urea, and TBE buffer, containing 89 mM Tris 37 °C for 3 h. base (pH 8.0), 89 mM boric acid, and 2 mM EDTA. The gel was run at 55 °C for around 1.5 h at 650 V on Hoefer SE 600 electrophoresis system and was stained with ethidium bromide (Sigma). The major band was cut under UV light and Native extraction of RNA from bacteria. After induction, centrifuge 1 ml bacteria eluted out. culture solution in a 1.5 ml centrifuge tube and remove the suspension. Re-suspend 8 NATURE COMMUNICATIONS (2018) 9:2196 DOI: 10.1038/s41467-018-04652-4 www.nature.com/naturecommunications | | | NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04652-4 ARTICLE the pellet in 100 µL Buffer L, containing 10 mM Tris-HCl (pH 7.4) and 10 mM Mg 25. Laughrea, M. & Jetté, L. Kissing-loop model of HIV-1 genome dimerization: (OAc) . Then destroy the bacterial membrane by adding 100 µL phenol solution HIV-1 RNAs can assume alternative dimeric forms, and all sequences (Sigma). Pipet out the aqueous layer and directly deposit it into native PAGE gel or upstream or downstream of hairpin 248− 271 are dispensable for dimer scan under AFM. To prepare samples for denaturing PAGE gel or gel puriﬁcation, formation. Biochemistry 35, 1589–1598 (1996). 10 µL NaOAc (3 M, pH 5.2) and 200 µL ethanol were added to 100 µL aqueous layer, 26. Guo, P., Erickson, S. & Anderson, D. A small viral RNA is required for in vitro followed by an ethanol participation in dry ice to get rid of salts. Alternatively, the cell packaging of bacteriophage phi 29 DNA. Science 236, 690–694 (1987). pellet was re-suspended in 15 ml Buffer L and sit in an ice-water bath. Cells were 27. Dibrov, S. M., Johnston-Cox, H., Weng, Y. & Hermann, T. Functional lysed by sonication (without phenol) with Branson Digital Soniﬁer (10% amplitude). architecture of HCV IRES domain II stabilized by divalent metal ions in the Sonicating for 5 s and stop 5 s; and repeat over until total of 10 min. 1 ml lysates were crystal and in solution. Ang. Chem. Int. Ed. 46, 226–229 (2007). centrifuged at 16,000×g for 30 min to remove the cell debris and the upper layer was 28. Seeman, N. C. DNA components for molecular architecture. Acc. Chem. Res. 2+ diluted 40 times with TAE/Mg buffer for AFM imaging. 30, 357–363 (1997). 29. Shih, W. M., Quispe, J. D. & Joyce, G. F. A 1.7-kilobase single-stranded DNA that folds into a nanoscale octahedron. Nature 427, 618–621 (2004). Data availability. Data supporting the ﬁndings of this study are available within the article and its Supplementary Information Files and from the corresponding 30. Ducani, C., Kaul, C., Moche, M., Shih, W. M. & Högberg, B. Enzymatic production of ‘monoclonal stoichiometric’ single-stranded DNA author upon reasonable request. oligonucleotides. Nat. Methods 10, 647–652 (2013). 31. Elbaz, J., Yin, P. & Voigt, C. A. Genetic encoding of DNA nanostructures and Received: 13 April 2018 Accepted: 14 May 2018 their self-assembly in living bacteria. Nat. Commun. 7, 11179 (2016). 32. Praetorius, F. et al. Biotechnological mass production of DNA origami. Nature 552,84–87 (2017). 33. Han, D. et al. Single-stranded DNA and RNA origami. Science 358, 1402 (2017). 34. Williams, S., Lund, K., Lin, C., Wonka, P. Lindsay, S. & Yan, H. Tiamat: a three-dimensional editing tool for complex DNA structures. In DNA References Computing. DNA 2008. Lecture Notes in Computer Science, Vol. 5347 (eds 1. Winfree, E., Liu, F., Wenzler, L. A. & Seeman, N. C. Design and self-assembly Goel, A. Simmel, F. C. & Sosik, P.) 90-101 (2008). of two-dimensional DNA crystals. Nature 394, 539–544 (1998). 35. Zuker, M. Mfold web server for nucleic acid folding and hybridization 2. Yan, H., Park, S. H., Finkelstein, G., Reif, J. H. & LaBean, T. H. DNA- prediction. Nucleic Acids Res. 31, 3406–3415 (2003). templated self-assembly of protein arrays and highly conductive nanowires. 36. Tang, G. et al. EMAN2: an extensible image processing suite for electron Science 301, 1882–1884 (2003). microscopy. J. Struct. Biol. 157,38–46 (2007). 3. He, Y. et al. Hierarchical self-assembly of DNA into symmetric 37. Pettersen, E. F. et al. UCSF Chimera—a visualization system for exploratory supramolecular polyhedra. Nature 452, 198–201 (2008). research and analysis. J. Comput. Chem. 25, 1605–1612 (2004). 4. Ke, Y., Ong, L. L., Shih, W. & Yi, P. Three-dimensional structures self- 38. Ponchon, L., Beauvais, G., Nonin-Lecomte, S. & Dardel, F. A generic protocol assembled from DNA bricks. Science 338, 1177–1183 (2012). for the expression and puriﬁcation of recombinant RNA in Escherichia coli 5. Rothemund, P. W. K. Folding DNA to create nanoscale shapes and patterns. using a tRNA scaffold. Nat. Protoc. 4, 947–959 (2009). Nature 440, 297–302 (2006). 39. Paul, E. & Cowtan, K. Coot: model-building tools for molecular graphics. Acta 6. Guo, P. The emerging ﬁeld of RNA nanotechnology. Nat. Nanotechnol. 5, Cryst. Sect. D: Biol. Cryst. 60, 2126–2132 (2004). 833–842 (2010). 7. Langecker, M. et al. Synthetic lipid membrane channels formed by designed Acknowledgements DNA nanostructures. Science 338, 932–936 (2012). We would like to thank NSF (CMMI - 1437301), ONR (N00014-15-1-2707), and NSFC 8. Chworos, A. et al. Building programmable jigsaw puzzles with RNA. Science (81429001) for ﬁnancial supports. 306, 2068–2072 (2004). 9. Geary, C., Rothemund, P. W. K. & Andersen, E. S. A single-stranded architecture for cotranscriptional folding of RNA nanostructures. Science 345, Author contributions 799–804 (2014). G.W. and C.M. conceived the project; C.M. supervised the project; M.L, M.Z., S.W., C.T., 10. Lo, P. K. et al. Loading and selective release of cargo in DNA nanotubes with and D.L. carried out the experiments; M.L, M.Z., S.W., C.T., D.L., Y.W., W.J., G.W., and longitudinal variation. Nat. Chem. 2, 319–328 (2010). C.M. analyzed the data, wrote, and commented on the manuscript. 11. Knudsen, J. B. et al. Routing of individual polymers in designed patterns. Nat. Nanotechnol. 10, 892–898 (2015). 12. Delebecque, C. J., Lindner, A. B., Silver, P. A. & Aldaye, F. A. Organization of Additional information intracellular reactions with rationally designed RNA assemblies. Science 333, Supplementary Information accompanies this paper at https://doi.org/10.1038/s41467- 470–474 (2011). 018-04652-4. 13. Afonin, K. A. et al. Design and self-assembly of siRNA-functionalized RNA nanoparticles for use in automated nanomedicine. Nat. Protoc. 6, 2022–2034 Competing interests: The authors declare no competing interests. (2011). 14. Ponchon, L. & Dardel, F. Recombinant RNA technology: the tRNA scaffold. Reprints and permission information is available online at http://npg.nature.com/ Nat. Methods 4, 571–576 (2007). reprintsandpermissions/ 15. Nelissen, F. H. T. et al. Fast production of homogeneous recombinant RNA – towards large-scale production of RNA. Nucleic Acids Res. 40,e102–e102 (2012). Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in 16. Farzadfard, F. & Lu, T. K. Genomically encoded analog memory with precise published maps and institutional afﬁliations. in vivo DNA writing in living cell populations. Science 346, 1256272 (2014). 17. Summers, J. & Mason, W. S. Replication of the genome of a hepatitis B-like virus by reverse transcription of an RNA intermediate. Cell 29, 403–415 (1982). 18. Lin, C. et al. In vivo cloning of artiﬁcial DNA nanostructures. Proc. Natl Acad. Open Access This article is licensed under a Creative Commons Sci. USA 105, 17626–17631 (2008). Attribution 4.0 International License, which permits use, sharing, 19. Whitelam, S. & Jack, R. L. The statistical mechanics of dynamic pathways to adaptation, distribution and reproduction in any medium or format, as long as you give self-assembly. Ann. Rev. Phys. Chem. 66, 143–163 (2015). appropriate credit to the original author(s) and the source, provide a link to the Creative 20. Mahen, E. M., Watson, P. Y., Cottrell, J. W. & Fedor, M. J. mRNA secondary Commons license, and indicate if changes were made. The images or other third party structures fold sequentially but exchange rapidly in vivo. PLoS Biol. 8, material in this article are included in the article’s Creative Commons license, unless e1000307 (2010). indicated otherwise in a credit line to the material. If material is not included in the 21. Mandal, M. & Breaker, R. R. Gene regulation by riboswitches. Nat. Rev. Mol. article’s Creative Commons license and your intended use is not permitted by statutory Cell Biol. 5, 451–463 (2004). regulation or exceeds the permitted use, you will need to obtain permission directly from 22. Svoboda, P. & Cara, A. D. Hairpin RNA: a secondary structure of primary the copyright holder. To view a copy of this license, visit http://creativecommons.org/ importance. Cell. Mol. Life Sci. 63, 901–908 (2006). licenses/by/4.0/. 23. Lescoute, A. & Westhof, E. Topology of three-way junctions in folded RNAs. RNA 12,83–93 (2006). © The Author(s) 2018 24. Shu, D., Shu, Y., Haque, F., Abdelmawla, S. & Guo, P. Thermodynamically stable RNA three-way junction for constructing multifunctional nanoparticles for delivery of therapeutics. Nat. Nanotechnol. 6, 658–667 (2011). NATURE COMMUNICATIONS (2018) 9:2196 DOI: 10.1038/s41467-018-04652-4 www.nature.com/naturecommunications 9 | | | http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nature Communications Springer Journals http://www.deepdyve.com/lp/springer-journals/in-vivo-production-of-rna-nanostructures-via-programmed-folding-of-9gj8sNXwa4

Loading next page...

Publisher: Springer Journals
Copyright: Copyright © 2018 by The Author(s)
Subject: Science, Humanities and Social Sciences, multidisciplinary; Science, Humanities and Social Sciences, multidisciplinary; Science, multidisciplinary
eISSN: 2041-1723
DOI: 10.1038/s41467-018-04652-4
Publisher site: See Article on Publisher Site

Abstract

ARTICLE DOI: 10.1038/s41467-018-04652-4 OPEN In vivo production of RNA nanostructures via programmed folding of single-stranded RNAs 1 1 1 1 2 2 3 Mo Li , Mengxi Zheng , Siyu Wu , Cheng Tian , Di Liu , Yossi Weizmann , Wen Jiang , 4 1 Guansong Wang & Chengde Mao Programmed self-assembly of nucleic acids is a powerful approach for nano-constructions. The assembled nanostructures have been explored for various applications. However, nucleic acid assembly often requires chemical or in vitro enzymatical synthesis of DNA or RNA, which is not a cost-effective production method on a large scale. In addition, the difﬁculty of cellular delivery limits the in vivo applications. Herein we report a strategy that mimics protein production. Gene-encoded DNA duplexes are transcribed into single-stranded RNAs, which self-fold into well-deﬁned RNA nanostructures in the same way as polypeptide chains fold into proteins. The resulting nanostructure contains only one component RNA molecule. This approach allows both in vitro and in vivo production of RNA nanostructures. In vivo synthesized RNA strands can fold into designed nanostructures inside cells. This work not only suggests a way to synthesize RNA nanostructures on a large scale and at a low cost but also facilitates the in vivo applications. 1 2 Department of Chemistry, Purdue University, West Lafayette, IN 47907, USA. Department of Chemistry, University of Chicago, Chicago, IL 60637, USA. 3 4 Markey Center for Structural Biology and Department of Biological Sciences, Purdue University, West Lafayette, IN 47907, USA. The Institute of Respiratory Diseases, Xinqiao Hospital, 400037 Chongqing, China. These authors contributed equally: Mo Li, Mengxi Zheng. Correspondence and requests for materials should be addressed to G.W. (email: wanggs2003@hotmail.com) or to C.M. (email: mao@purdue.edu) NATURE COMMUNICATIONS (2018) 9:2196 DOI: 10.1038/s41467-018-04652-4 www.nature.com/naturecommunications 1 | | | 1234567890():,; ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04652-4 ucleic acids (DNA and RNA) have been extensively C) to a low temperature (e.g., 25 °C) over a long period of time. explored for molecular self-assembly and a wide range of Obviously, this thermal annealing process is not feasible for 1– Nnanostructures have been assembled from nucleic acids. nucleic acid self-folding in vivo. A potential approach is to design Such nanostructures can be applied to various ﬁelds from the targeted nucleic acid nanostructures both thermodynamically 10–13 physical devices to biomedical applications. Following DNA stable and kinetically favorable. To achieve this goal, the ssRNA is nanotechnology, programmed RNA self-assembly has rapidly designed to fold following a sequential and hierarchical pathway. evolved in hope that RNA has more structural complexity and Newly synthesized ssRNA would ﬁrst fold into hairpins while functional diversity. Up to now, DNA/RNA self-assembly gen- transcription. Hairpin structures are not only thermodynamically erally starts from chemically or enzymatically synthesized stable but also topologically simple. They only involve local single-stranded DNA or RNA (ssDNA or ssRNA). This method is interactions, thus, fold quickly. If any alternative structure forms, not desirable for large-scale production because of the excessive it would readily rearrange into the target hairpin structure via cost. A potential solution is to clone nucleic acids in bacteria, such local branch migration. Upon hairpin formation, which deﬁnes 14, 15 as E. coli, in the same way as recombinant proteins. How- the RNA’s secondary structure, most of the RNA residues are ever, DNA molecules normally exist as duplexes in the cell. inert as being in the content of duplexes, leaving a minimal Though direct cellular production of ssDNAs is possible amount of RNA residues as unpaired. The unpaired residues are 16, 17 in some special cases such as by reverse transcriptase or M13 able to further form long-range tertiary interactions, leading to bacterial genomes , there are limitations. For instance, the formation of fully folded, designed nanostructures. The the ssDNA length range is limited or additional enzymatic overall folding pathway is similar to that of the naturally occur- 18 21 treatment after puriﬁcation is needed. Comparatively, a more ring complex RNA structures, such as hairpin ribozymes. . suitable choice is RNA. Cellular RNAs primarily exist as single- Conceptually, the design concept resembles the principle that stranded, and their length can vary in a broad range. Recently, developed by Geary et al. However, a signiﬁcant change is that the Geary et al. have demonstrated that ssRNAs, in test-tubes, can short dovetail seams (2–3 bps) are avoided. Such short helical cotranscriptionally fold into designed nanomotifs, which can domains are not very stable and are likely to deformation under further assemble into 2D arrays. Delebecque et al. have cloned mild stress. RNA nanostructures in E. coli to organize chemical reactions in vivo. However, the nanoscaled, structural details of the RNA complexes have not been thoroughly characterized under native Results conditions. Molecular design. The RNA nanostructures in this study are In this work, we have developed a versatile strategy to prepare rationally designed based on natural RNA motifs and tertiary well-deﬁned nanostructures by folding individual long ssRNAs. interactions (Fig. 1), including: (i) RNA duplexes, (ii) RNA 22 23 Each nanostructure contains only one ssRNA molecule. The hairpins , (iii) 3-way junctions in open conformation (o3WJ) , resulting nanostructures can be cloned, expressed, and self-folded (vi) 3-way junctions in stacked conformation (s3WJ) observed in in E. coli. RNA nanostructures have been thoroughly character- the packaging RNA (pRNA) of phi29 bacteriophage, (v) ized by gel electrophoresis, atomic force microscopy (AFM) coaxially stacked kissing loops (KLs) found in the dimerization imaging, and cryogenic electron microscopy (cryoEM). initiation sites of HIV-1 RNA, (vi) a 3-way loop (3WL) inter- A key challenge of this approach is to design the folding action observed in phi29 pRNA, (vii) 4-way junctions in open pathway to avoid kinetic traps. For nucleic acid self-assembly, the conformation (o4WJ), and (viii) 90°–kink found in the internal target structures are designed to be thermodynamically stable, but ribosome entry site (IRES) of the hepatitis C virus (HCV) RNA often not kinetically favored. This problem is commonly solved genome. According to the molecular designs, DNA templates by slowly cooling the samples from a high temperature (e.g., 95 ° coding for the ssRNAs have been synthesized so that RNAs can Duplex Hairpin 90°-kink Kissing-loops (KL) Open 3-way junction (o3WJ) Open 4-way junction (o4WJ) Stacked 3-way junction (s3WJ) 3-way loops (3WL) Fig. 1 Component motifs of RNA structures. For each motif, a schematic drawing and a 3D model are shown. The thick colored lines and thin gray lines represent RNA backbones and basepairs, respectively 2 NATURE COMMUNICATIONS (2018) 9:2196 DOI: 10.1038/s41467-018-04652-4 www.nature.com/naturecommunications | | | ~10 nm ~10 nm NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04652-4 ARTICLE be produced by in vitro transcription. Alternatively, the DNA RNA square and double-square.We ﬁrst tested this approach by templates can be inserted into a plasmid and introduced into E. the assembly of two 2D structures: a square (S) and a double- coli. Then cellular transcription machinery will transcribe the square (S2). Figure 2 shows the designs and experimental char- gene into the corresponding ssRNA, which spontaneously folds acterizations (For RNA sequences and 3D models, see Supple- into the designed nanostructures inside the cells. mentary Figs 1 & 2). RNA S contains four 90°-kinks and a KL 3′ 2+ Mg 5′ 3′ 2+ Mg 5′ S2 ~17.5 nm cd S* S* 300 bp 220 bp 100 bp S2 300 bp 220 bp S2 100 bp Fig. 2 Designs, folding, and characterization of an RNA square (S) and an RNA double square (S2). a, b The molecular design and single-stranded folding pathways for S and S2, respectively. The RNA single strands are colored in a rainbow gradient from 5′ (red) to 3′ end (purple). Red, green, and blue boxes with dashed lines highlight a 90°-kink, a KL interaction, and a 3WL interaction, respectively. Each edge is composed of a two-turn RNA duplex. Characterization of c, d RNA square and e, f RNA double square. c, e electrophoretic analysis; d, f Atomic force microscopy (AFM) imaging. Note that S* has the same sequence as S except that one loop sequence is altered so that no KL interaction can form. (Scale bar: 20 nm) NATURE COMMUNICATIONS (2018) 9:2196 DOI: 10.1038/s41467-018-04652-4 www.nature.com/naturecommunications 3 | | | DNA markers DNA markers Quenched Quenched S* Annealed S Annealed As transcribed S* As S* transcribed ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04652-4 interacting pair, while S2 contains four 90°-kinks and two 3WL potential of expressing RNA nanostructures inside cells. It is pairs. After in vitro transcription, puriﬁed RNA molecules were worthy noting that for single-component RNA folding method, thermally annealed (slowly cooled from 65 °C to 4 °C over 2 h in a any experimental issue related to stoichiometry is avoided. 2+ neutral, Mg -containing, aqueous buffer) to promote the RNA Direct visualization by AFM conﬁrms that the RNA molecules 2+ folding. RNA molecules ﬁrst folded into Mg –independent, indeed fold into the designed structures (Fig. 2d, f). In the AFM intermediate, loose secondary structures, which then further fol- images, randomly distributed RNA particles exhibited uniform 2+ ded into Mg –dependent, compact, tertiary structures via loop- sizes. Their apparent shapes are closed and consistent with the loop (KL or 3WL) interactions. The resulting RNA square (S) designs: S particles as squares and S2 particles as double squares. samples were analyzed by native PAGE (Fig. 2c). A dominant In contrast, the control molecule, S*, appears as a very different, sharp band is observed; suggesting that the RNA square (S)is open geometry because the KL interaction is disrupted. Please fully folded. To show the critical role of the KL interaction in the note that AFM images give a direct visualization of the RNA folding, a control molecule (S*) is prepared, in which one loop structures. However, discrete frameworks could be easily sequence is altered so that the KL interaction is interrupted disrupted during sample preparation and by AFM probes, (Supplementary Fig. 1). S* migrates slightly slower than S, indi- resulting in artiﬁcial structural heterogeneity. Thus, the RNA cating that S* can fold into the open intermediate secondary folding efﬁciency should not be estimated by AFM imaging of a structure, but not the compact tertiary structure. The formation small number of objects; instead, native PAGE of the bulk sample of S is a unimolecular process and is expected to be kinetically gives a reliable estimation. fast. To prove this hypothesis, we performed a quenching experiment by plunging the RNA solutions from 65 °C onto the RNA nanoﬂower and tetra-square. To demonstrate that this ice. In the native PAGE, the quenched samples (both S and S*) approach is versatile and could be used to assemble large, com- migrate identically to those annealed samples, indicating that the plex geometry, a 5-petal nanoﬂower and a tetra-square (S4) were intramolecular folding of the designed RNA molecules is indeed a designed (Fig. 3, Supplementary Figs 3 and 4). The RNA ﬂower is fast process. In addition, the comparison between S and S* 1571 bases long and the ﬁnal structure contains ﬁve petals. The molecules gives an estimation of the folding yield of S close to folding pathway is illustrated in Supplementary Fig. 3b. The RNA 100% as no RNA in the lane of S migrates like the partially folded quickly folded into an intermediate, highly branched, secondary S*. More strikingly, the RNA molecules could spontaneously fold structure, which is composed of multiple duplex regions and into the designed structures cotranscriptionally. After transcrip- multiple loops. Via KL interactions, the branch ends pairwisely tion, the crude RNA samples were directly analyzed by PAGE associated with each other to form the ﬂower petals. Note that ten without any puriﬁcation or thermal treatment (the lane as indi- 90°-kink loops, six KL interactions, and ten 3WLs taken from cated As Transcribed). The migration pattern was identical to phi29-pRNA were integrated into the RNA strand to facilitate the those of puriﬁed and thermally treated samples, indicating that, as desired folding. The expected ﬁve-petal-ﬂower particles were expected, the RNAs folded into the designed structures even in clearly observed under AFM (Fig. 3b), conﬁrming that the RNA crude solutions without any annealing step. A similar result was indeed folded into the designed structure. In the transcription also observed for S2 (Fig. 2e). This result points to a great mixture from a crude PCR mixture (no puriﬁcation for DNA ab Transcription mixture Purified RNA cd Transcription mixture Purified RNA Fig. 3 Folding of complex RNA nanostructures from single RNA strands. a Structural design of an RNA 5-petal ﬂower, which contains six KL interactions, ten 90°-kinks, and ten 3WJs. b AFM images of the RNA ﬂowers. Left: transcription mixture from a PCR mixture. Right: thermal annealed, puriﬁed RNA nanoﬂower molecule. Scale of inset: 60 nm. c Structural design of an RNA tetra-square (S4), which contains four 3WLs, four 90°-kinks, and a 4WJ. d AFM images of the RNA S4. Left: transcription mixture. Right: thermal annealed, puriﬁed RNA S4. Scale of inset: 25 nm. (Scale bar: 50 nm) 4 NATURE COMMUNICATIONS (2018) 9:2196 DOI: 10.1038/s41467-018-04652-4 www.nature.com/naturecommunications | | | NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04652-4 ARTICLE Fig. 4 Folding of a single-stranded, 4-turn, RNA tetrahedron (T4). a Structural design of T4. An RNA single strand is rainbow colored from 5′ to 3′ end. It contains four 3WJs and will fold from a 2D branching structure 5′ into a 3D tetrahedron upon the three pairs of KL interactions (indicated by 3′ KL T4 dashed, double-arrowed lines). Each edge is four helical turns long. Scale Interaction bar, 50 nm. b A control molecule T4*. One loop sequence is altered to prevent one KL interaction. Thus T4* will assemble into a ﬂat, double- triangular shape instead of a tetrahedron. The models were built with 39 37 Coot and Chimera . All KLs are colored orange in the model. Scale bar, 50 nm. c, d AFM images of T4 and T4*, respectively. For each structure, three particles are zoomed-in and ﬁtted with corresponding shapes. e, g 5′ Cryogenic electron microscopy (cryoEM) characterization of T4. e A raw 3′ KL T4* cryoEM image. Each white box indicates an individual RNA particle. Scale Interaction bar, 50 nm. f Four different views of the reconstructed structural model of the RNA tetrahedron (top) and corresponding views of the simulated model (bottom). Scale bar, 5 nm. g Pairwise comparison between raw cryoEM images of individual particles (left) and the corresponding projections (right) of the reconstructed structural model. The raw particles were T4 selected from different images to represent views at different orientations fold with high yields. And this strategy was further conﬁrmed by the success of the folding of an RNA tetra-square (S4) structure (Fig. 3c, d and Supplementary Fig. 4). RNA tetrahedron. One important test of molecular self-assembly is to generate discrete, 3D nanostructures, which can be readily T4* achieved by the reported strategy (Fig. 4). For demonstration, a tetrahedral structure (T4) is designed to fold from a 623-base- 2 long ssRNA. It contains six edges and four vertices. All edges are four helical turns long. Three of them are standard A-form duplexes and each of the other three contains a KL interaction in the middle. Each of the vertexes is an o3WJ containing four unpaired uracils on each strand at the center to ensure sufﬁcient out-of-plane ﬂexibility to fold. Upon cooling, the ssRNA ﬁrst folded into a three-branched structure (Fig. 4a, left), and then closed into a tetrahedral geometry via KL interactions (Fig. 4a, right). To show the importance of the KL interactions, a control molecule (T4*) was prepared, in which a pair of KL interaction was disrupted by sequence mutation. It is expected to only fold into an open, 2D, double-triangular shape instead of the compact, fully folded, 3D tetrahedron (Fig. 4b). We have characterized the tetrahedron by native PAGE (Supplementary Fig. 6), AFM imaging (Fig. 4c, d), and cryoEM (Fig. 4e–g). PAGE analysis suggests that RNA molecules efﬁciently fold into the designed structures (Supplementary Fig. 6a). Each sample appears as a sharp, dominant band in the corresponding lane. The migration pattern is consistent with the expected structures. The fully-folded, compact, 3D tetrahedron (T4) has a higher mobility than the open, 2D, double-triangle structure (T4*). The folding is fast, and no difference has been observed between quenching and annealing for each sample. In the lane of T4, no RNA migrates like the partially folded T4* molecule; suggesting that all T4 molecules are fully folded. In order to conﬁrm the migration pattern and the folding ability, we have prepared two more RNA tetrahedrons with different sizes templates and RNA), some linear structures (150–400 nm long) and their corresponding control molecules: a three-turn-edged were observed. They were unfolded or partially folded RNA tetrahedron (T3 and T3*, 484 bases long, Supplementary Fig. 5) molecules, or truncated RNA molecules and DNA templates (the and a two-turn-edged tetrahedron (T2 and T2*, 340 bases long, designed DNA template is 1591 bps or 540 nm long). In the Supplementary Fig. 5). Similar migration patterns are observed puriﬁed RNA samples, only the 5-petal nanoﬂowers remained in (Supplementary Fig. 6b-c). AFM imaging gives more direct the AFM image. Such a complicated pattern would be quite structural clues. T4 samples appear compact and consistent with challenging for the traditional tile-based method, which would a collapsed tetrahedral geometry (Fig. 4c), while T4* samples involve multiple different tiles and sophisticated inter-tile inter- exhibit clear double-triangle structures (Fig. 4d and Supplemen- actions, generally leading to a low assembly yield. With this tary Fig. 7). Please note that some T4 structures are disturbed ssRNA folding strategy, complicated structures could quickly self- during AFM imaging. NATURE COMMUNICATIONS (2018) 9:2196 DOI: 10.1038/s41467-018-04652-4 www.nature.com/naturecommunications 5 | | | ~11 nm ~11 nm ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04652-4 CryoEM study revealed the native, 3D, tetrahedral structure of IPTG), abundant double-square shaped particles are clearly T4. In raw images of the T4 sample (Fig. 4e), uniform-sized identiﬁed, which are identical to the S2 sample prepared from particles could be easily identiﬁed (indicated with white boxes). in vitro transcribed RNA (Fig. 5c: samples from phenol lysis Their shapes resembled the 2D projections of a tetrahedron. method; Supplementary Fig. 9d: samples from sonication From the observed individual particles, a technique of single- method). In contrast, no double-square shaped particles are particle 3D reconstruction was applied and resulted in a observed from the cells without the S2 gene or the cells without structural model of a tetrahedron at a resolution of 4.1 nm expression induction. Based on the band intensities in the PAGE (Fig. 4f). To verify the reconstructed model, two comparisons graph (Supplementary Fig. 9b), we can estimate that S2 molecule were performed: the computed 2D projections with individual accounts for roughly 11.2% of all RNAs in the cell; suggesting that raw particles (Fig. 4g) and with class averages of raw particles in vivo expression is potentially an efﬁcient way to produce RNA with similar views (see Supplementary Fig. 8). Clear similarities in nanostructures. Similarly, we can in vivo express the tetra-square the pairwise comparisons veriﬁed the reconstructed model. Note (S4) structure (Supplementary Fig. 10). It seems that the RNA that the six struts in the model were not the same. folding is efﬁcient, but the RNA transcription yield is low or the RNA is not stable in the cell. In the future, it might be worthy using natural RNAs, e.g., tRNA or 5 S rRNA, as scaffolds to In vivo production of RNA nanostructures. The most promis- improve the RNA production yields of longer and complex ing feature of the current approach is that it is compatible with 14, 15 strands. in vivo expression of the designed RNA nanostructures. In our design, each RNA nanostructure contains one single strand. Therefore, an RNA strand, in the same way as peptides fold into Discussion proteins, can spontaneously fold into the designed structure In vivo production of nucleic acid (DNA/RNA) nanostructures is under the physiological condition during in vivo transcription. To a major challenging to the ﬁeld of nucleic acid nanotechnology demonstrate this capability, we have cloned and expressed the and has attracted signiﬁcant efforts. The original idea was ﬁrst double square structure (S2)in E. coli (Fig. 5). Brieﬂy, the S2- proposed by Nadrian Seeman in 1997. But the complicated coding DNA sequence is inserted into a pET23a plasmid, a topology associated with that design prevented it from experi- bacterial expression vector, under the control of a T7 promoter mental realization. The ﬁrst experimental progress was elegantly (Supplementary Fig. 9a). The resulting recombinant plasmid is done by Shih et al. in 2004. They assembled an octahedron from then transformed into BL21 (DE3) E. coli cells and the RNA a long enzymatically generated DNA single strand and a few transcription is induced by isopropyl β-D-1-thiogalactopyranoside helper strands. In 2006, Paul Rothemund developed a powerful (IPTG). Then the cells were lysed by phenol or sonication. and generally applicable strategy, so-called DNA origami. It fol- Without any further manipulation, the aqueous solution of the ded long, biologically produced, M13 bacteriophage genome by total cell lysates was directly examined by both PAGE and AFM hundreds of chemically synthesized, short DNA strands into imaging. On native PAGE, the desired RNA molecule from E. coli designed nanostructures. In 2008, Lin et al. cleverly used bacteria migrates the same as the in vitro prepared S2 sample (Supple- to produce simple DNA nanostructures, such as a DNA four way mentary Fig. 9c), suggesting that the RNA from the cells is well junction. In 2011, Delebecque et al. produced RNA molecules, folded as the expected double-squared structure. This is further which further associated with each other to organize chemical conﬁrmed by AFM imaging (Fig. 5c and Supplementary Fig. 9d). reactions in E. coli. In 2011, Afonin et al. reported in vitro, In the raw cell lysate after expression induction (addition of co-transcriptional assembly of multi-stranded RNA complexes. a b ~10 nm E. coli – Plasmid – IPTG + IPTG Fig. 5 Cloning and in vivo expression of RNA nanostructures. a Overall scheme. An expression vector (pET23a, orange) carrying an S2-coding DNA sequence (green) is transformed into E. coli cells. Upon IPTG induction, the S2 gene gets transcribed into RNAs, which self-fold into the designed double- square structure in the cell. b Structural model of the S2 structure. c AFM imaging of cell lysates from E. coli without the expression vector (- Plasmid), or with the vector but without S2 gene expression induction (- IPTG), or with S2 gene expression ( + IPTG). (Scale bar: 20 nm) 6 NATURE COMMUNICATIONS (2018) 9:2196 DOI: 10.1038/s41467-018-04652-4 www.nature.com/naturecommunications | | | ~17.5 nm NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04652-4 ARTICLE Table 1 Sequence of DNA templates S5′-TTCTAATACGACTCACTATACCGTGTACTCGGAGGAACTACTCATAT TCGAGCGTCGGAGGAACTACTGCAGTGGCTCGTTATTCGG AACTACTATGGTGAAGGAGGCACGCCATAGCCGAATAACGAGCCATTGCAGCCTCCGACGCTCGAATGTGAGCCTCCGAGTACACGGGCTG AGCCACGTGAAGCCTCCACGCGTGGAACTACTCAGCAGTGCTAAGAATCTA-3′ S* Identical to S except the underlined sequence was replaced by AAAAAAAAA S2 5′-TTCTAATACGACTCACTATAGACTCGGAGGAACTACTCACATTC GAGCGTTGATGGAGGTCAACGCTCGAATGTGAGCCTCTGAGTCCT GTCGATCAGCCTGGTCTAGCATCTATCAGCCCAGATTCAACTATTGGGGATACTCCCTGAATCGGATCACGCACGGAACTACTGTAGTGCTT CGTTAGTCGGAACTACTGCTCTAGATAGGTTGATGTCGGTCAACCTATCTAGAGCAGCCGACTGACGAAGCATTACAGCCGTGTGTGATCCG ATTCAGCCCAGTAGTTGAATCTGGGGATACGACCTGATAGGTGCTAGGCCAGGAACTACTGATTGACAGGCAGTGCTAAGAATCTA-3′ FL 5′-TTCTAATACGACTCACTATACTCTAGAAGACCGAGGCTGGCTATAT GAGGAACTACTCGGTGAAGAGGAGACGCCGAGCCTCATGTAGCCAGTCTCGGTCTCCGTCGTCAGACACTCGGGTGTGTGTCGTTGCGTCG ACGGGTTGTGTGTCCTGTCTTGGCAGGTCTGGACCGGGAGCGCTTGAGCTCAGCCATTTGAACTCCTCACGAATGGAACTACTGAGTTCAA GTGCTCCTGGTCCGGACCTGTCAAGACGGGACACTTTGTAGGGTGGTGTTTCACGGGACATCACCTGTGGAGGGTTATCGGTACTCTGGG CACGGAACTACTGCATGAATCGAGCACGTGCAGCCGTGCTCAGAGTGCCGATGACCCTCTACAGGTGGTGTCCCGTGTGTAACGTCTCGAA GAGGAGGTGTGTCAAGTAGGCACAGCTGAGGTTCGAACGGACTTCGGCTCCCAGCCTGGTGAAGCTCGAACGCCAGGAACTACTGGGAG TCGAAGTCTGTTCGAATCTCAGTTGTGCCTGCTTGACTTTGTCAGGGAGTGTTTCAGTAGTAACGGGCCGACTAACTGGAGCGTGGGAAGA TCATTGGAACTACTAGCTGAAACGGCAACGGCTAGCCAATGGTCTTCCTACGCTCTAGTTGGTCGGTCCGTTGCTACTGTGTGTTTGAACAC TGGACGAACACTACTCCCTGATCACCTTCTCTTTGAGATGTTACTACCACCCTATCAACCTGTCGGCGCAATGACACTACCCTGGGATCAGT CTGATCTATCGTGACGACTAGAGCTCGATCAACAGGCTATTCGGTCAAGACTACTCTCAGATCACGGTGAACCAGTGACGCCGTGTGTCGA GTCAACTGCGTTGCGTCCGTATCACTACCTCTTGGAGAGCCCAGTGAATGCCGTACGCTGGGAACTACTCTCCGAGAGGTGGTGATGCGGA TGCAACGTAGTTGGCTCGACTTTGTCTGAGAGTGTTTCAGCATGCGTCAGCTCTGTGGTGAATTCAGCTAGCACTGTCAGGAACTACTCAGT GAAGATAGGACGCTGAGCCTGACGGTGCTAGTTGAATTCATCACAGGGCTGATGCATGCTGTGTCTTGGCCGAGTAGCTTGTTGTGTCGGT GTCACCGTATGGACATCTGTCTTCAGCATTGGCCCTAGCCGGCTGAACCTATCACGGCCGGAACTACTAGGGTCAATGCTGGAGACAGGTG TCCATGCGGTGGCACCGACTTTGTCGAGCTCTGTTTCACTAGGAGGATCCGAGCATTCTTCACTGGGGTTCTCGTGGTGGAACTACTCACCG AAGCACGTACGGTGAGCCACCATGAGAACTCCAGTGGAGAATGTTCGGATTCTCCTAGTGTGTCGTCGCGATGGATCGGACTGTGTGCCTG GACGCAGTCCCTTACGCTCTGCTAGGGTGCTACTTAGCCAAGTGAAACGTGCACGCTTGGAACTACTAAGTGGCACCTTAGCGGAGCGTGA GGGATTGCGTTCAGGCACTTTGTCCCAGGGTGTTTCACCCGGGTGTCTGATGACGGACTAGCATAACCCCTTGGGGCCTCTAAACGGG-3′ S4 5′-TTCTAATACGACTCACTATAGGTCTAGAGGTGAGTGTGTAACTACG GTCTAGTCGTGGAGGAACTGGTCGTAGTTCCTCCGCGACTAGGCCGACACACTCACCCATCACTCCAGCCACTTCCAGGGAAACTCCGGCTT CGTGTGATCTCGACATTCTGGGGATACGACCTGGTTACCCTCCTAGACTATCAACTAGGTGGGAGTGAATGCTGTTCTGGGAGTAGAACAGC GTTCACTCTCACCGATAGTCTGGGAGGGTAATCAGCCCAGAGTGTCGAGATTACACGTTCGTACAGACACGTCTGAGTGGGGATACTCCCT GAACTCAGGTGCTTGAGCTCAACTAGGAGTTGAGTCTTCTCAACGTGTTGCTACGTTGAGAAGATTCAACTCCGAGTTCAAGCACCTGGGTT CAGCCCACTCAGGCGTGTCTGTGCGTTGTGAGTATCGACAGGATAGAGGGATAGCAACGTGCAGTGGAGAACATTACACAACTAGTGTCAG AGCTAGCAGGTCGTGAGCGTACGACCTGCTGGCTCTGACACGTGTGATGTTCTCCATTGCACGCCTCTGTCCTGTCGATGCTCACTTGCTG GAGTTTCTCTGGAAGTGGGATACGCTCTGGAGTGATGCCTTGGGGCCTCTAAACGGG-3′ T4 5′-TTCTAATACGACTCACTATAGACGTGTAGTTGTCTGTTACCTTTTTA GCGTGTAGATAGTCGCTACAAAAGCGGTATGTAGCGACTATCTACGCGCTTTTTAATCAGACGCAAGTTTATCAACACCTGAGATAGACTTG CGTTTGATTTTTTAGGTAACAGACAACTACATGTCCCCAATCTAGTCTCAGTCCGGATTTTGAGGATGAGCACATTTGCGCGTATGTGGTCC GAAGGTCATGGCATTTTTCCGTACATCAACTTAGGTGCAACAGGTGAGCATCTAAGTTGATGTACGGATTTTGTGCGTCGCTCGACGGCCGAA ACGGCAACGGTCGTCGAGCGATGCACTTTTTGCCATGACCTTCGGACCACATACGCGCGAATGTGCTTATCCTCTTTTTCATGCGATGGCT CGGCGGCTCTTGGTACATTCTGGCTCGACTTTTTTAGGCTTCTCTGTCACGGATGCAATGCCGTAGCATTCGTGACAGAGAAGCCTTTTTAG GTTAACTACACGGATATAAACCGCTAATATCTGTGTAGTTAGCCTTTTTAAGTCGAGCCAGAATGTACCAAGAGCCGTCGAGCCATTGCATG ATTTTTCCGGATTGAGACTGGATTGGGGCAGTGCTAAGAATCTA -3′ T4* Identical to T4 except the underlined sequence was replaced by AATTTTTTA T3 5′-TTCTAATACGACTCACTATAGACGTGTAGTCGCCTGTTACCTTTTTA GCGGCTACAAAAGCGGTATGTAGCCGCTTTTTAATCAGACGCAAGTTTATCAACACCTGAGATAAACTTGCGTCTGATTTTTTAGGTAACA GGCGACTACACGTCCCCAATCCGGATTTTGAGGATGAGCACATTCGCGCGTAGGTCATGGCATTTTTCCGTACTGCAACAGGTGAGCAGTAC GGATTTTGTGCATCGCTCGACGGCCGAAACGGCAACGGCCGTCGAGCGATGCACTTTTTGCCATGACCTACGCGCGAATGTGCTCATCCT CTTTTTCATGCAATGGCTCGGTAATTCTGGCTCGACTTTTTTAGGCTAATGCAATGCCGTAGCATTAGCCTTTTTAGGCTAACTACACGGAT ATAAACCGCTAATATCCGTGTAGTTAGCCTTTTTAAGTCGAGCCAGAATTACCGAGCCATTGCGTGATTTTTCTGGGTTGGG-3′ T3* Identical to T3 except the underlined sequence was replaced by AATTTTTTA T2 5′-TTCTAATACGACTCACTATAGGCTACTGTTTATAGCGAGGCCGTG ATTCGCTTTTTAAAAGTCATAATGCCGTAATGACTTTTTTTTTCGA GCCGAATCGCCAACGGCTCGAATTTAAGCGAATCACGGCCTCGCTATTTTAATCGATGCATCTGACTGCTCTTTTAGACGCUCGAATGGCGA ACGGGCGTCTTTTAGTACGTCCAAGTCCACAGGACGTACTTTTAGAGCAGTCAGATGCATCGATTTTTCAGTAGCCGCATTGUATACTGCTT TCTGAGTACCAAGTGGACAGGTACTTAGTTTCGGTGCGGCAAACGGCAAGCCGCACCGTTTGCAGTATGCAGTGCTAAGAATCTA-3′ T2* Identical to T2 except the underlined sequence was replaced by AATTTTTTA S2’ 5′-TTCTAATACGACTCACTATACTCTAGAACAGGCGGAGGAACTACTC ACGTTCCAGCGTTCATGGAGGTGAACGCTGGAATGTGAGCCTCTGCCTGCTCCTGATCAGCCACGTGTGACATCTAAGAGCCCATCCTCACC AGTAGGGGATACTCCCTGAATCGGAACATGGACGGAACTACTGTAGTGCTTGGTTAGTCGGAACTACTGCACGAGGTAGGTTGATGTCGGTC AACCTATCTCGTGCAGCCGACTGACCAAGCATTACAGCCGTCCGTGTTCCGATTCAGCCCTGCTGGTGAGGATGGGGATACGACCTCTTAG GTGTCACGCGTGGAACTACTGATTAGGAGCTAGCATAACCCCTTGGGGCCTCTAAACGGG-3′ T7 promoter sequences are highlighted in bold Two restriction enzyme sites, XbaI and StyI are formatted in italics (for in vivo double-square S2′ and for tetra-square S4) (FL: RNA nanoﬂower, S2′: double-square for in vivo expression) In 2013, Ducani et al. enzymatically, either in vitro or in vivo, in vitro fold long RNA strands into nanostructures. This strategy produced ssDNA, which were released with restriction enzymes is closely related to our work. In 2016, Elbaz et al. took advantage in test-tubes and used to assemble DNA nanostructures with of reverse transcriptase to in vivo produce ssDNAs, which could 30 31 thermal annealing. In 2014, Geary et al. developed a strategy to in vivo assemble into multi-stranded DNA nanostructures. In NATURE COMMUNICATIONS (2018) 9:2196 DOI: 10.1038/s41467-018-04652-4 www.nature.com/naturecommunications 7 | | | ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04652-4 2017, Hendrik and coworkers used bacteria to cost-effectively and Native PAGE. 4% native PAGE gel was prepared with 19:1 acrylamide//bisacry- 2+ lamide gel and TAE/Mg buffer. The gel was run at 4 °C. Then stained with at large-scale produce single DNA strands, which in vitro self- Stains-All (Sigma) and scanned by an HP scanner (Scanjet 4070 Photosmart). assembled into multi-stranded, DNA nanostructures upon ther- mal annealing. In 2017, Yan and coworkers enzymatically Agarose gel analysis. Samples were loaded into 1% agarose gel and run in TBE produced DNA or RNA single strands, which, upon thermal buffer on a FB-SB-710 electrophoresis unit (FisherBiotech) at room temperature annealing, in vitro folded into designed nanostructures without under constant voltage of 80 V. The gel was stained in EB and illuminated under topological tangling. UV light. Images were taken by cell phone. In conclusion, we have developed a programmable ssRNA folding strategy for the preparation of complex nanostructures. AFM imaging. Mica surface (Ted Pella, Inc.) was pre-treated with poly-L-lysine (0.1% w/v, Ted Pella, Inc.) to increase the adsorption of small RNA particles. 20 µL Each nanostructure is primarily composed of RNA duplexes and poly-L-lysine (5 µg/ml) was added onto freshly cleaved mica substrate and was is folded from one long ssRNA. Thus, the formation of the incubated for 30 s. Then the surface was washed by 100 µL H O and dried by nanostructures is independent of stoichiometry and RNA con- compressed air. Annealed RNA solution was diluted to 50 nM and a drop of 5 µL 2+ centration. This strategy allows cloning and expression of RNA was deposited onto pre-treated mica and incubated for 2 min. 20 µL TAE/Mg nanostructures inside of cells in the same way as those of buffer was further added. Imaging was performed in a ﬂuid cell under tapping mode on a Multimode AFM (MMAFM 2, Veeco) with SNL-10 probe (Veeco, Inc.). recombinant proteins; pointing to a cost-effective way for large- scale production of nucleic acid nanostructures. We expect this Cryo-EM imaging.3 μL RNA solution (50 nM) was deposited on the Lacey strategy would facilitate a wide range of in vivo biomedical Carbon-only grid (300 mesh Cu, Ted Pella, Inc.). The grid was blotted with two applications of RNA nanotechnology, such as integrating multiple pieces of ﬁlter paper (Whatman, 1001-055) and was immediately ﬂash frozen in functional RNA moieties to regulate cellular processes. liquid nitrogen-cooled liquid ethane. Images were taken on FEI Titan Krios with − 2 accelerating voltage of 300 kV under 90 e /Ǻ dose condition at under-defocus 8 μm to enhance image contrast. Methods Oligonucleotides. Double-stranded DNA templates were purchased from IDT, Inc. Polymerase chain reaction was applied to amplify DNA duplexes into large Single particle reconstruction. The single particle reconstruction was carried out amount. The sequence of DNA templates can be found in Table 1. using computer software EMAN 2. Totally, 470 particles were selected to go Sequence of PCR primers: through reconstruction process using EMAN2 reconstruction package, following P1: 5′-TTCTAATACGACTCACTATA-3′ the standard protocol. (Wiki at: http://blake.bcm.edu). A general summary of the P2: 5′-TAGATTCTTAGCACTGC-3′ steps includes: (1) Import and convert CCD micrographs; (2) Pick typical particles; P3: 5′-CCCGTTTAGAGGCC-3′ (3) CTF determination; (4) Remove bad particles and build particle sets; (5) Make P4: 5′-CGTTATTGTACCCAACCCAGAAAAATCAC-3′ 2D reference-free class-averages; (6) Build an initial 3D model; (7) 3D high reso- S/S1*/S2/T2/T2*/T4/T4*: P1 + P2; FL/S2’/S4: P1 + P3; T3/T3*: P1 + P4. lution reﬁnement; (8) Evaluate the resolution of 3D map. About 470 randomly selected particles were used to build 12 class averages, which were used to generate initial models under different symmetries, such as C3 symmetry, TET symmetry or DNA sequence design. When designing RNA sequences, we take two parts into without symmetry (see Supplementary Fig. 8). Totally around 360 particles were consideration. (i) The single stranded loops in the RNA secondary structures, used for 3D reﬁnements of the RNA tetrahedron. The reﬁnement was carried out including the 90°-kink loop, KLs, pRNA interacting loops. They are critical regions with a 2° angle interval under C3 symmetry. A projection matching algorithm was and determine the long-range interactions and local conformations. For the 5-nt- applied for the determination of the center and orientation of raw particles in long, 90°-kink loop, the sequence is strictly taken from the HCV IRES domain II. iterative reﬁnement. After reﬁnement, the class-averages of raw particles and the For KLs and pRNA loops, the exact sequence is not important, but sequence 2D projections of 3D model were generated in two ﬁles but paired in order. The complementarities are ensured for the interacting pairs. (ii) The remaining duplex comparisons shown in the result were manually selected to present the distinct and regions. They provide all the struts in the structures and do not involve long-range comprehensive orientations. The resolution of resulted structural density map was interactions. Their sequences are randomly generated by a computer software determined to be at 4.1 nm. Final reconstruction results were visualized by UCSF Tiamat. Roughly 10% of the Watson-Crick basepairs (A–T and G–C) are replaced 37 Chimera software . by G–U pairs to facilitate PCR ampliﬁcation of the DNA templates. Such G–U pairs are evenly distributed along the duplex regions. Finally, the RNA sequences are checked by a computer software Mfold to make sure that the RNA will fold into Recombinant plasmid preparation. Double-stranded DNA template and pET23a plasmid (GenScript Inc.) were double-digested by XbaI and StyI-HF restriction the designed secondary structures. If there are any undesired, strong secondary structures, the sequence will be manually altered to avoid such structures. enzymes (New England BioLabs Inc.) by following the manufacturer- recommended protocol. The reaction was performed in 100 µL CutSmart buffer at 37 °C for 2 h, and 20 units of each enzyme were added. After digestion, DNA was Polymerase chain reaction. DNA double-stranded template was ampliﬁed by recovered by DNA clean & concentrator columns (Zymo Research Corp.) to get rid polymerase chain reaction using Taq DNA polymerase kit (New England Biolabs of proteins. Then the template and plasmid were ligated by T4 DNA ligase (New Inc.). 4–10 ng DNA template, 200 µM dNTPs, 200 nM of forward and reverse England BioLabs Inc.), and were incubated overnight at 16 °C. primers were dissolved in 100 µL standard Taq reaction buffer. Then 0.5 µL Taq DNA polymerase was added to the system. The solution was initially denatured for Plasmid cloning and bacteria culture. Ligation product was combined with XL1- 30 s at 95 °C, and started 30 cycles at 95 °C for 30 s, 50 °C for 1 min, and 68 °C for Blue competent Escherichia coli (E. coli) cells for transformation. The mixture was 2–3 min. Then kept the ﬁnal extension at 68 °C for another 5 min. PCR product kept on ice for 30 min, and then heat to 42 °C for 90 s and put on ice for 2 min. was stored at 4 °C for short time storage and at −20 °C for long terms. Spread the transformed bacteria on Lysogeny Broth (LB, 100 µg/ml ampicillin was added) (RPI. Corp.) plates and incubate overnight at 37 °C. After the colonies In vitro RNA preparation. RNA molecules were synthesized from in vitro tran- growing into desired size, inoculate a colony into an individual aliquot of 6 ml LB/ scription using T7 RNA polymerase (AmpliScribe T7-Flash transcription kit; ampicillin liquid medium and shake overnight at 37 °C. 1.5 ml bacteria solution was Epicenter, Inc.). Corresponding DNA template was added, and the experiment was taken out into another 6 ml fresh LB/ampicillin liquid medium and kept shaking at conducted by following the manufacturer-recommended protocol. 37 °C until UV absorbance was around 0.5 OD at 600 nm, measuring by a UV/Vis spectrophotometer (Beckman coulter, DU520). Plasmid was extracted using plas- mid isolation kit (Zymo Research Corp.). The puriﬁed plasmid was characterized by 2+ Formation of ssRNA structures. ssRNA was prepared in TAE/Mg buffer and restriction enzyme digestion and gel analysis to conﬁrm the correct insert size. annealed by different processes: Quenched---65 °C/5 min, 0 °C/5 min; Regular anneal---65 °C/5 min, 50 °C/30 min, 37 °C/30 min, 22 °C/30 min, and 4 °C/30 min. 2+ TAE/Mg buffer contained 40 mM Tris base (pH 8.0), 20 mM acetic acid, 2 mM In vivo RNA expression induction. The recombinant plasmid was transformed TM EDTA, and 12.5 mM magnesium acetate. into BL21 Star (DE3) competent E.coli cells (Life Technologies Corp.) for RNA expression, following the manufacturer’s protocol. The cell culture process was the same as described above. When UV absorbance was around 0.5 OD at 600 nm, Denaturing PAGE. 5% denaturing PAGE gel was prepared with the 19:1 acryla- IPTG (1 mM) was added for induction and the solution was shaken at mide/bisacrylamide solution, 8 M urea, and TBE buffer, containing 89 mM Tris 37 °C for 3 h. base (pH 8.0), 89 mM boric acid, and 2 mM EDTA. The gel was run at 55 °C for around 1.5 h at 650 V on Hoefer SE 600 electrophoresis system and was stained with ethidium bromide (Sigma). The major band was cut under UV light and Native extraction of RNA from bacteria. After induction, centrifuge 1 ml bacteria eluted out. culture solution in a 1.5 ml centrifuge tube and remove the suspension. Re-suspend 8 NATURE COMMUNICATIONS (2018) 9:2196 DOI: 10.1038/s41467-018-04652-4 www.nature.com/naturecommunications | | | NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04652-4 ARTICLE the pellet in 100 µL Buffer L, containing 10 mM Tris-HCl (pH 7.4) and 10 mM Mg 25. Laughrea, M. & Jetté, L. Kissing-loop model of HIV-1 genome dimerization: (OAc) . Then destroy the bacterial membrane by adding 100 µL phenol solution HIV-1 RNAs can assume alternative dimeric forms, and all sequences (Sigma). Pipet out the aqueous layer and directly deposit it into native PAGE gel or upstream or downstream of hairpin 248− 271 are dispensable for dimer scan under AFM. To prepare samples for denaturing PAGE gel or gel puriﬁcation, formation. Biochemistry 35, 1589–1598 (1996). 10 µL NaOAc (3 M, pH 5.2) and 200 µL ethanol were added to 100 µL aqueous layer, 26. Guo, P., Erickson, S. & Anderson, D. A small viral RNA is required for in vitro followed by an ethanol participation in dry ice to get rid of salts. Alternatively, the cell packaging of bacteriophage phi 29 DNA. Science 236, 690–694 (1987). pellet was re-suspended in 15 ml Buffer L and sit in an ice-water bath. Cells were 27. Dibrov, S. M., Johnston-Cox, H., Weng, Y. & Hermann, T. Functional lysed by sonication (without phenol) with Branson Digital Soniﬁer (10% amplitude). architecture of HCV IRES domain II stabilized by divalent metal ions in the Sonicating for 5 s and stop 5 s; and repeat over until total of 10 min. 1 ml lysates were crystal and in solution. Ang. Chem. Int. Ed. 46, 226–229 (2007). centrifuged at 16,000×g for 30 min to remove the cell debris and the upper layer was 28. Seeman, N. C. DNA components for molecular architecture. Acc. Chem. Res. 2+ diluted 40 times with TAE/Mg buffer for AFM imaging. 30, 357–363 (1997). 29. Shih, W. M., Quispe, J. D. & Joyce, G. F. A 1.7-kilobase single-stranded DNA that folds into a nanoscale octahedron. Nature 427, 618–621 (2004). Data availability. Data supporting the ﬁndings of this study are available within the article and its Supplementary Information Files and from the corresponding 30. Ducani, C., Kaul, C., Moche, M., Shih, W. M. & Högberg, B. Enzymatic production of ‘monoclonal stoichiometric’ single-stranded DNA author upon reasonable request. oligonucleotides. Nat. Methods 10, 647–652 (2013). 31. Elbaz, J., Yin, P. & Voigt, C. A. Genetic encoding of DNA nanostructures and Received: 13 April 2018 Accepted: 14 May 2018 their self-assembly in living bacteria. Nat. Commun. 7, 11179 (2016). 32. Praetorius, F. et al. Biotechnological mass production of DNA origami. Nature 552,84–87 (2017). 33. Han, D. et al. Single-stranded DNA and RNA origami. Science 358, 1402 (2017). 34. Williams, S., Lund, K., Lin, C., Wonka, P. Lindsay, S. & Yan, H. Tiamat: a three-dimensional editing tool for complex DNA structures. In DNA References Computing. DNA 2008. Lecture Notes in Computer Science, Vol. 5347 (eds 1. Winfree, E., Liu, F., Wenzler, L. A. & Seeman, N. C. Design and self-assembly Goel, A. Simmel, F. C. & Sosik, P.) 90-101 (2008). of two-dimensional DNA crystals. Nature 394, 539–544 (1998). 35. Zuker, M. Mfold web server for nucleic acid folding and hybridization 2. Yan, H., Park, S. H., Finkelstein, G., Reif, J. H. & LaBean, T. H. DNA- prediction. Nucleic Acids Res. 31, 3406–3415 (2003). templated self-assembly of protein arrays and highly conductive nanowires. 36. Tang, G. et al. EMAN2: an extensible image processing suite for electron Science 301, 1882–1884 (2003). microscopy. J. Struct. Biol. 157,38–46 (2007). 3. He, Y. et al. Hierarchical self-assembly of DNA into symmetric 37. Pettersen, E. F. et al. UCSF Chimera—a visualization system for exploratory supramolecular polyhedra. Nature 452, 198–201 (2008). research and analysis. J. Comput. Chem. 25, 1605–1612 (2004). 4. Ke, Y., Ong, L. L., Shih, W. & Yi, P. Three-dimensional structures self- 38. Ponchon, L., Beauvais, G., Nonin-Lecomte, S. & Dardel, F. A generic protocol assembled from DNA bricks. Science 338, 1177–1183 (2012). for the expression and puriﬁcation of recombinant RNA in Escherichia coli 5. Rothemund, P. W. K. Folding DNA to create nanoscale shapes and patterns. using a tRNA scaffold. Nat. Protoc. 4, 947–959 (2009). Nature 440, 297–302 (2006). 39. Paul, E. & Cowtan, K. Coot: model-building tools for molecular graphics. Acta 6. Guo, P. The emerging ﬁeld of RNA nanotechnology. Nat. Nanotechnol. 5, Cryst. Sect. D: Biol. Cryst. 60, 2126–2132 (2004). 833–842 (2010). 7. Langecker, M. et al. Synthetic lipid membrane channels formed by designed Acknowledgements DNA nanostructures. Science 338, 932–936 (2012). We would like to thank NSF (CMMI - 1437301), ONR (N00014-15-1-2707), and NSFC 8. Chworos, A. et al. Building programmable jigsaw puzzles with RNA. Science (81429001) for ﬁnancial supports. 306, 2068–2072 (2004). 9. Geary, C., Rothemund, P. W. K. & Andersen, E. S. A single-stranded architecture for cotranscriptional folding of RNA nanostructures. Science 345, Author contributions 799–804 (2014). G.W. and C.M. conceived the project; C.M. supervised the project; M.L, M.Z., S.W., C.T., 10. Lo, P. K. et al. Loading and selective release of cargo in DNA nanotubes with and D.L. carried out the experiments; M.L, M.Z., S.W., C.T., D.L., Y.W., W.J., G.W., and longitudinal variation. Nat. Chem. 2, 319–328 (2010). C.M. analyzed the data, wrote, and commented on the manuscript. 11. Knudsen, J. B. et al. Routing of individual polymers in designed patterns. Nat. Nanotechnol. 10, 892–898 (2015). 12. Delebecque, C. J., Lindner, A. B., Silver, P. A. & Aldaye, F. A. Organization of Additional information intracellular reactions with rationally designed RNA assemblies. Science 333, Supplementary Information accompanies this paper at https://doi.org/10.1038/s41467- 470–474 (2011). 018-04652-4. 13. Afonin, K. A. et al. Design and self-assembly of siRNA-functionalized RNA nanoparticles for use in automated nanomedicine. Nat. Protoc. 6, 2022–2034 Competing interests: The authors declare no competing interests. (2011). 14. Ponchon, L. & Dardel, F. Recombinant RNA technology: the tRNA scaffold. Reprints and permission information is available online at http://npg.nature.com/ Nat. Methods 4, 571–576 (2007). reprintsandpermissions/ 15. Nelissen, F. H. T. et al. Fast production of homogeneous recombinant RNA – towards large-scale production of RNA. Nucleic Acids Res. 40,e102–e102 (2012). Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in 16. Farzadfard, F. & Lu, T. K. Genomically encoded analog memory with precise published maps and institutional afﬁliations. in vivo DNA writing in living cell populations. Science 346, 1256272 (2014). 17. Summers, J. & Mason, W. S. Replication of the genome of a hepatitis B-like virus by reverse transcription of an RNA intermediate. Cell 29, 403–415 (1982). 18. Lin, C. et al. In vivo cloning of artiﬁcial DNA nanostructures. Proc. Natl Acad. Open Access This article is licensed under a Creative Commons Sci. USA 105, 17626–17631 (2008). Attribution 4.0 International License, which permits use, sharing, 19. Whitelam, S. & Jack, R. L. The statistical mechanics of dynamic pathways to adaptation, distribution and reproduction in any medium or format, as long as you give self-assembly. Ann. Rev. Phys. Chem. 66, 143–163 (2015). appropriate credit to the original author(s) and the source, provide a link to the Creative 20. Mahen, E. M., Watson, P. Y., Cottrell, J. W. & Fedor, M. J. mRNA secondary Commons license, and indicate if changes were made. The images or other third party structures fold sequentially but exchange rapidly in vivo. PLoS Biol. 8, material in this article are included in the article’s Creative Commons license, unless e1000307 (2010). indicated otherwise in a credit line to the material. If material is not included in the 21. Mandal, M. & Breaker, R. R. Gene regulation by riboswitches. Nat. Rev. Mol. article’s Creative Commons license and your intended use is not permitted by statutory Cell Biol. 5, 451–463 (2004). regulation or exceeds the permitted use, you will need to obtain permission directly from 22. Svoboda, P. & Cara, A. D. Hairpin RNA: a secondary structure of primary the copyright holder. To view a copy of this license, visit http://creativecommons.org/ importance. Cell. Mol. Life Sci. 63, 901–908 (2006). licenses/by/4.0/. 23. Lescoute, A. & Westhof, E. Topology of three-way junctions in folded RNAs. RNA 12,83–93 (2006). © The Author(s) 2018 24. Shu, D., Shu, Y., Haque, F., Abdelmawla, S. & Guo, P. Thermodynamically stable RNA three-way junction for constructing multifunctional nanoparticles for delivery of therapeutics. Nat. Nanotechnol. 6, 658–667 (2011). NATURE COMMUNICATIONS (2018) 9:2196 DOI: 10.1038/s41467-018-04652-4 www.nature.com/naturecommunications 9 | | |

Journal

Nature Communications – Springer Journals

Published: Jun 6, 2018

References

Design and self-assembly of two-dimensional DNA crystals

Winfree, E; Liu, F; Wenzler, LA; Seeman, NC
DNA-templated self-assembly of protein arrays and highly conductive nanowires

Yan, H; Park, SH; Finkelstein, G; Reif, JH; LaBean, TH
Hierarchical self-assembly of DNA into symmetric supramolecular polyhedra

He, Y
Three-dimensional structures self-assembled from DNA bricks

Ke, Y; Ong, LL; Shih, W; Yi, P
Folding DNA to create nanoscale shapes and patterns

Rothemund, PWK
The emerging field of RNA nanotechnology

Guo, P
Synthetic lipid membrane channels formed by designed DNA nanostructures

Langecker, M
Building programmable jigsaw puzzles with RNA

Chworos, A
A single-stranded architecture for cotranscriptional folding of RNA nanostructures

Geary, C; Rothemund, PWK; Andersen, ES
Loading and selective release of cargo in DNA nanotubes with longitudinal variation

Lo, PK
Routing of individual polymers in designed patterns

Knudsen, JB
Organization of intracellular reactions with rationally designed RNA assemblies

Delebecque, CJ; Lindner, AB; Silver, PA; Aldaye, FA
Design and self-assembly of siRNA-functionalized RNA nanoparticles for use in automated nanomedicine

Afonin, KA
Recombinant RNA technology: the tRNA scaffold

Ponchon, L; Dardel, F
Fast production of homogeneous recombinant RNA – towards large-scale production of RNA

Nelissen, FHT
Genomically encoded analog memory with precise in vivo DNA writing in living cell populations

Farzadfard, F; Lu, TK
Replication of the genome of a hepatitis B-like virus by reverse transcription of an RNA intermediate

Summers, J; Mason, WS
In vivo cloning of artificial DNA nanostructures

Lin, C
The statistical mechanics of dynamic pathways to self-assembly

Whitelam, S; Jack, RL
mRNA secondary structures fold sequentially but exchange rapidly in vivo

Mahen, EM; Watson, PY; Cottrell, JW; Fedor, MJ
Gene regulation by riboswitches

Mandal, M; Breaker, RR
Hairpin RNA: a secondary structure of primary importance

Svoboda, P; Cara, AD
Topology of three-way junctions in folded RNAs

Lescoute, A; Westhof, E
Thermodynamically stable RNA three-way junction for constructing multifunctional nanoparticles for delivery of therapeutics

Shu, D; Shu, Y; Haque, F; Abdelmawla, S; Guo, P
Kissing-loop model of HIV-1 genome dimerization: HIV-1 RNAs can assume alternative dimeric forms, and all sequences upstream or downstream of hairpin 248− 271 are dispensable for dimer formation

Laughrea, M; Jetté, L
A small viral RNA is required for in vitro packaging of bacteriophage phi 29 DNA

Guo, P; Erickson, S; Anderson, D
Functional architecture of HCV IRES domain II stabilized by divalent metal ions in the crystal and in solution

Dibrov, SM; Johnston-Cox, H; Weng, Y; Hermann, T
DNA components for molecular architecture

Seeman, NC
A 1.7-kilobase single-stranded DNA that folds into a nanoscale octahedron

Shih, WM; Quispe, JD; Joyce, GF
Enzymatic production of ‘monoclonal stoichiometric’ single-stranded DNA oligonucleotides

Ducani, C; Kaul, C; Moche, M; Shih, WM; Högberg, B
Genetic encoding of DNA nanostructures and their self-assembly in living bacteria

Elbaz, J; Yin, P; Voigt, CA
Biotechnological mass production of DNA origami

Praetorius, F
Single-stranded DNA and RNA origami

Han, D
Mfold web server for nucleic acid folding and hybridization prediction

Zuker, M
EMAN2: an extensible image processing suite for electron microscopy

Tang, G
UCSF Chimera—a visualization system for exploratory research and analysis

Pettersen, EF
A generic protocol for the expression and purification of recombinant RNA in Escherichia coli using a tRNA scaffold

Ponchon, L; Beauvais, G; Nonin-Lecomte, S; Dardel, F
Coot: model-building tools for molecular graphics

Paul, E; Cowtan, K

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

In vivo production of RNA nanostructures via programmed folding of single-stranded RNAs

In vivo production of RNA nanostructures via programmed folding of single-stranded RNAs

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

In vivo production of RNA nanostructures via programmed folding of single-stranded RNAs

In vivo production of RNA nanostructures via programmed folding of single-stranded RNAs

Abstract

Journal

Recommended Articles

References

Our policy towards the use of cookies