RNA ORIGAMI DESIGN

INTRODUCTION

In 2014, Geary and his colleagues first reported the beauty of RNA origami nanostructures (1). For our NanoWolves team project, we will elevate the impact of RNA origami by using and demonstrating RNA origami as a functional biomolecule for therapeutic applications. At the beginning stage, we need to understand basic concepts of RNA origami design.

Figure 1.  (A) Illustration of RNA origami designs. (B) AFM images of RNA lattices made of RNA origami. Images obtained from Geary et. al., 2014, (1)

Figure 1. (A) Illustration of RNA origami designs. (B) AFM images of RNA lattices made of RNA origami. Images obtained from Geary et. al., 2014, (1)

This page provides a brief overview of designing RNA origami. There are four main steps in designing RNA origami: (i) creating a 3D model, (ii) conversion to a 2D model, (iii) producing a text file design from the 2D model, and (iv) design/analysis of the RNA sequence. It is important to know that the final design will be composed of a single RNA strand that folds back upon itself into a complex 3D shape. For this reason the end result of this overview will consist of a single RNA sequence.

BACKGROUND OF RNA ORIGAMI DESIGN

For the design of RNA origami, we use an A-form of RNA double helix (Figure 2). The geometry of the helix is important since it will help determine where to place the strand’s crossover points. We use multiple structural motifs in this design. The use of a kissing loop motif allows for a single strand to be routed through a multi-helical structure. In this example we use a 180° kissing loop. The tetra loop is a small four-nucleotide motif that caps the ends of the structure. Tetra loops increase the structure’s stability.
The design process of RNA origami can be divided into four main parts: (i) create 3D model, (ii) convert to 2D model, (iii) sequence design, and (iv) sequence analysis. More detailed information on RNA design can be found in an RNA origami design tutorial reported by Sparvath, S. L et al. 2017 (2).  

GO TO TOP

CREATE 3D MODEL

In this project, we created a 2-helix RNA origami tethered with thrombin RNA aptamers.  Here we explain the main concepts and workflow of RNA origami design.
To start you obtain all parts (PDB files) of the origami structure. The required files are the RNA double helix, tetra loop, 180° Kissing loop, and the RNA thrombin aptamer (if available). The first step is to obtain an A-form RNA double helix, which will serve as a foundation for all other motifs (Figure 2) . To obtain the helix go to the make-na server (http://structure.usc.edu/makena/server.html). 

Figure 2  View of A-form RNA double helix in  Chimera .

Figure 2 View of A-form RNA double helix in Chimera.

Next, we obtain a 3D structure of the kissing-loop from the RNA junction website (https://rnajunc-tion.ncifcrf.gov). Kissing-loop with the angle range of 175-185. (no.13070) was downloaded and saved in PDB Format (text). Figure 3 show the 3D structure of 180°kissing loop in UCSF Chimera.

Figure 3  View of 180 °kissing loop (PDB:2B8R) in Chimera.

Figure 3 View of 180 °kissing loop (PDB:2B8R) in Chimera.

We obtained the tetra loop (PDB: 1F7Y) from the RCSB PDB website (www.rcsb.org/). This file contains the tetra loop embedded in a larger structure, which needed to be extracted. To extract the tetra loop we used Swiss-PdbViewer, which can be downloaded from http://spdbv.vital-it.ch. The sequence from G30 to U39 was extracted from the large structure as shown in Figure 4.

Figure 4  Tetra loop (PDB:1F7Y) before (A) and after (B) extraction from larger structure. The red circle shows tetra loop within the larger structure.

Figure 4 Tetra loop (PDB:1F7Y) before (A) and after (B) extraction from larger structure. The red circle shows tetra loop within the larger structure.

The last file needed is the exosite-binding RNA aptamer. We obtain the exosite-II RNA aptamer (PDB: 3DD2) from the RCSB PDB website (www.rcsb.org/). The 3D structure of exosite-I RNA aptamer does not exist. The exosite-II aptamer needs to be extracted from a larger structure (Figure 5). We also used Swiss-PdbViewer for extraction of desired parts.

Figure 5.  Exosite II RNA Thrombin aptamer (PDB: 3DD2) before (A) and after (B) extraction from larger structure. The red circle shows RNA aptamer within the larger structure.

Figure 5. Exosite II RNA Thrombin aptamer (PDB: 3DD2) before (A) and after (B) extraction from larger structure. The red circle shows RNA aptamer within the larger structure.

All motif files have been obtained. Next, all motifs are assembled and aligned properly by using Chimera, which can be downloaded from www.cgl.ucsf.edu/chimera. First, we align two A-form RNA helices as shown in Figure 6. Next, we color the nucleotides where the crossovers will occur, this provides a better visualization of where the crossovers are to be placed.

Figure 6  Two double helices after being properly aligned to form crossovers. Top helix is Helix 1. Bottom helix is Helix 2. The phosphate atoms and backbone and side (sugar/base) where the crossovers take place are identified in red and yellow, respectively.

Figure 6 Two double helices after being properly aligned to form crossovers. Top helix is Helix 1. Bottom helix is Helix 2. The phosphate atoms and backbone and side (sugar/base) where the crossovers take place are identified in red and yellow, respectively.

After RNA helices were aligned, all motifs including tetra-loops, kissing-loops and RNA aptamer were inserted into the alignment panel. The alignment process was done manually using the command line. The alignment of all motif is shown in Figure 7

Figure 7  Side view of RNA structure with all motifs aligned with the helices. Top left motif is the RNA thrombin aptamer.

Figure 7 Side view of RNA structure with all motifs aligned with the helices. Top left motif is the RNA thrombin aptamer.

The next step is to convert the structure to a single stranded structure. To do so you need to acquire the ligate.pl script from the Ebbe Anderson’s lab website (www.andersen-lab.dk). The ligation structure of RNA origami is shown in Figure 8

Figure 8  The 3D model of 2-helix RNA origami containing an RNA aptamer.

Figure 8 The 3D model of 2-helix RNA origami containing an RNA aptamer.

2) CREATE 2D MODEL

Next, convert the RNA 3D model to a 2D model by simply opening the assembled RNA origami file, from the previous step, in the Assemble2 software.  The 2D model result is shown in Figure 9 (top panel). Then, the 2D model is transcribe into a text file (Figure 9, bottom). This part is completed manually.

Figure 9  Transcription from the 2D model to the text file. (Top) 2D illustration of RNA origami produced by Assemble2. (Bottom) Manual creation of 2D text file that is further used for sequence design by NUPACK. Note that the two designs do not look identical. Some creative visualization is required to produce the text file design.

Figure 9 Transcription from the 2D model to the text file. (Top) 2D illustration of RNA origami produced by Assemble2. (Bottom) Manual creation of 2D text file that is further used for sequence design by NUPACK. Note that the two designs do not look identical. Some creative visualization is required to produce the text file design.

3) SEQUENCE DESIGN

To generate an RNA sequence use NUPACK online software (www.nupack.org). The 2D text file from the previous step is then run through a trace script that can be found in Ebbe Anderson’s lab website (www.andersen-lab.dk). The output code from the trace script is shown in Figure 10. The output RNA sequences generated by NUPACK are shown in Figure 11.

Figure 10  The text file after running it through the trace script. The red box highlights the output code that can be submitted to NUPACK, which then provides the RNA sequence.

Figure 10 The text file after running it through the trace script. The red box highlights the output code that can be submitted to NUPACK, which then provides the RNA sequence.

Figure 11.  RNA sequences provided by  NUPACK . The  NED  (normalized ensemble defect) is also provided, usually the sequence with the lowest  NED  is chosen for analysis.

Figure 11. RNA sequences provided by NUPACK. The NED (normalized ensemble defect) is also provided, usually the sequence with the lowest NED is chosen for analysis.

4) SEQUENCE ANALYSIS

The last step is to analyze the RNA sequences from NUPACK. To test for a properly folding RNA origami, we used NUPACK and Mfold (link). The sequence can be analyzed to select an optimal sequence that assures proper structure folding. A good sequence is one with a low ∆G, a GC% of less than 65%, and a low normalized ensemble defect (NED). The fewer alternative secondary structures that a sequence has the better. To optimize the sequence, we manually edit sequences such as changing position of G-C base pairing in order to remove undesired structures.  

go to top

PLACEMENT OF RNA APTAMERS ON RNA ORIGAMI

Figure 12  The 2D folding simulation of thrombin RNA aptamers. (A) RNAR9D-14T,and (B) Toggle-25 aptamer (3,4)

Figure 12 The 2D folding simulation of thrombin RNA aptamers. (A) RNAR9D-14T,and (B) Toggle-25 aptamer (3,4)

RNA origami decorated with two RNA aptamers that specifically bind with exosite-I and -II of thrombin was designed. We selected two RNA aptamers: (i) RNAR9D-14T binds to prothrombin and thrombin at exosite-I (3), and (ii) Toggle-25 binds to exosite-II of thrombin (4). The RNAR9D-14T and Toggle-25 aptamer will be further called “apt-1 and -2”, respectively. The simulated folding structures of apt-1 and -2 are shown in Figure 12. The 3D motif of Apt-2 (PDB: 3DD2) can be found on Protein Data Bank while the 3D structure of apt-1 has not been solved. Therefore, we are only able to create a 3D model of RNA origami showing apt-2. For the sequence design, we used the 2D model of RNA origami as a core structure for appending apt-1 and -2. 

To improve and optimize anticoagulation activity, we placed two aptamers on RNA origami in four different positions. A two-helix RNA origami (2HO-RNA-XXXX) offers four logical positions for attaching RNA aptamers as shown in Figure 13.

Figure 13  (A) The 3D illustration and (B) 2D ribbon model of 2-helices RNA origami without RNA aptamers, 2HO-RNA-NNNN. The numbers represent four possible positions to tether RNA aptamer on RNA origami. Tetra-loops (aptamer attachment sites) represented in yellow boxes and the kissing loop is indicated in a green box

Figure 13 (A) The 3D illustration and (B) 2D ribbon model of 2-helices RNA origami without RNA aptamers, 2HO-RNA-NNNN. The numbers represent four possible positions to tether RNA aptamer on RNA origami. Tetra-loops (aptamer attachment sites) represented in yellow boxes and the kissing loop is indicated in a green box

We hypothesized that the binding activity of dual RNA aptamers tethered to RNA origami will depend upon the distance between the two aptamers and the flexibility of the aptamers at the various positions on the origami. Therefore, we designed four configurations of RNA aptamer placement on RNA origami as shown in Figure 14. For naming the four designs, we added four digits after RNA origami system such as 2HO-RNA-XXXX. Depending on the aptamer and placement position, X was replaced by aptamer name. The exosite I- and exosite II- binding RNA aptamers are called “1” and “2”, respectively. No aptamer is defined as “N”.  For example, 2HO-RNA-1N2N refers to exosite I-binding aptamer placed at position 1 and  exosite II-binding aptamer 2 tethered at position 3 on a 2-helix origami. All four designs are depicted in Figure 14. The 2HO-RNA-NNNN was used as a negative control in functional assays.

Figure 14  The 2D models of all four designs of RNA origami tethering two aptamers. (A) 2HO-RNA-NNNN. (B) 2HO-RNA-12NN. (3) 2HO-RNA-1N2N and (4) 2HO-RNA-2NN1

Figure 14 The 2D models of all four designs of RNA origami tethering two aptamers. (A) 2HO-RNA-NNNN. (B) 2HO-RNA-12NN. (3) 2HO-RNA-1N2N and (4) 2HO-RNA-2NN1

COMPUTATIONAL ANALYSIS OF RNA ORIGAI FOLDING

go to top

The folding of RNA origami structures was computationally analyzed by mfold (http://unafold.rna.albany.edu)  and NUPACK (http://www.nupack.org) online software. The kissing loop formation is not shown in the simulated structures. The folding results from mfold and NUPACK agreed with the 2D model as shown in Figures 15-18

Figure 15  (A) The 2D model of 2HO-RNA-NNNN and computational simulation of RNA folding analyzed by (B) mfold and (C) NUPACK. Tetra loops and kissing loop are indicated in yellow and green boxes, respectively.

Figure 15 (A) The 2D model of 2HO-RNA-NNNN and computational simulation of RNA folding analyzed by (B) mfold and (C) NUPACK. Tetra loops and kissing loop are indicated in yellow and green boxes, respectively.

Figure 16  (A) The 2D model of 2HO-RNA-12NN and computational simulation of RNA folding analyzed by (B) mfold and (C) NUPACK. Tetra loops and kissing loop are indicated in yellow and green boxes, respectively. Exosite 1 and 2-binding RNA aptamers are represented in purple and blue rectangles, respectively

Figure 16 (A) The 2D model of 2HO-RNA-12NN and computational simulation of RNA folding analyzed by (B) mfold and (C) NUPACK. Tetra loops and kissing loop are indicated in yellow and green boxes, respectively. Exosite 1 and 2-binding RNA aptamers are represented in purple and blue rectangles, respectively

Figure 17  (A) The 2D model of 2HO-RNA-1N2N and computational simulation of RNA folding analyzed by (B) mfold and (C) NUPACK. Tetre loops and kissing loop are indicated in yellow and green boxes, respectively. Exosite 1 and 2-binding RNA aptamers are represented in purple and blue rectangles, respectively.

Figure 17 (A) The 2D model of 2HO-RNA-1N2N and computational simulation of RNA folding analyzed by (B) mfold and (C) NUPACK. Tetre loops and kissing loop are indicated in yellow and green boxes, respectively. Exosite 1 and 2-binding RNA aptamers are represented in purple and blue rectangles, respectively.

Figure 18  (A) The 2D model of 2HO-RNA-2NN1 and computational simulation of RNA folding analyzed by (B) mfold and (C) NUPACK. Tetre loops and kissing loop are indicated in yellow and green boxes, respectively. Exosite 1 and 2-binding RNA aptamers are represented in purple and blue rectangles, respectively.

Figure 18 (A) The 2D model of 2HO-RNA-2NN1 and computational simulation of RNA folding analyzed by (B) mfold and (C) NUPACK. Tetre loops and kissing loop are indicated in yellow and green boxes, respectively. Exosite 1 and 2-binding RNA aptamers are represented in purple and blue rectangles, respectively.

SEQUENCE DESIGN FOR IN VITRO PRODUCTION

RNA origami structures without aptamers are 210 nucleotides in length, and 260 nucleotides in length for two aptamer tethered RNA origami. Chemical synthesis of long RNA (over 120 nucleotides) is difficult with current knowledge and technology. In cells, long coding RNA sequences, longer than 800 nucleotides, are normally produced. The solution for synthesizing long RNA strand is to employ biomachinary in the transcription process. We decided to use T7 RNA polymerase for RNA production via in vitro and in vivo transcription because of it’s well-known properties. For transcription process, we used double-stranded DNA as a template. The DNA template contains the Blueprint for RNA origami as well as the T7 promoter located at the 5’ end of the sequence (Figure 19).

Figure 19  The RNA origami production by  in vitro  transcription

Figure 19 The RNA origami production by in vitro transcription

CONCLUSION

We described the whole process of RNA origami design including extraction of motifs, alignment of pieces, creation of 3D model, construction of 2D model, sequence design, and structure analysis. We successfully designed a two-helical RNA origami that was used as a core structure to tether thrombin RNA aptamers. The 3D and 2D model provide valuable information relating to the structure of the RNA design. The 3D model provides a better view of the RNA structure, allowing one to detect any spatial problems that can lead to improper folding or hinder the functionality of the motif. The 2D model provides a clear outline of what the text file should look like, making the creation of the text file easier. These are just some examples of how the 3D and 2D model can be helpful. It is also important to know that there is no assurance of proper RNA folding by using computational analysis. Even so, one should further analyze all sequences, using the tools provided by NUPACK and Mfold, to select the sequence that has the greatest possibility of providing proper folding.

go to top

REFEREENCES

  1. Geary, C., Rothemund, P. W., and Andersen, E. S.,A single-stranded architecture for cotranscriptional folding of RNA nanostructures, Science. 2014, 345, 799-804.

  2. Sparvath, S. L., Geary. C. W., and Andersen, E. S., Computer-Aided Design of RNA Origami Structures, Methods Mol Biol. 2017, 1500, 51-80.

  3. Bompiani, K. M., Monroe, D. M., Church, F. C., and Sullenger B. A., A high affinity, antidote-controllable prothrombin, and thrombin-binding RNA aptamer inhibits thrombin generation and thrombin activity, J Thromb Haemost. 2012, 10, :870-80

  4. Long, S. B., Long, M. B., White, R. R., Sullenger, B. A., Crystal structure of an RNA aptamer bound to thrombin, RNA. 2008, 14, 2504-12.

go to top