Composing RNA Nanostructures from a Syntax of RNA Structural Modules

Abstract

Natural stable RNAs fold and assemble into complex three-dimensional architectures by relying on the hierarchical formation of intricate, recurrent networks of noncovalent tertiary interactions. These sequence-dependent networks specify RNA structural modules enabling orientational and topological control of helical struts to form larger self-folding domains. Borrowing concepts from linguistics, we defined an extended structural syntax of RNA modules for programming RNA strands to assemble into complex, responsive nanostructures under both thermodynamic and kinetic control. Based on this syntax, various RNA building blocks promote the multimolecular assembly of objects with well-defined three-dimensional shapes as well as the isothermal folding of long RNAs into complex single-stranded nanostructures during transcription. This work offers a glimpse of the limitless potential of RNA as an informational medium for designing programmable and functional nanomaterials useful for synthetic biology, nanomedicine, and nanotechnology.

Keywords: RNA architectonics; RNA folding; RNA nanotechnology; RNA self-assembly; nanoparticles; nanostructures; tectoRNAs.

Figure 1.

RNA structural syntax behind the rational design of 3D RNA nanostructures. (left) RNA 3D modules defining angles ranging from 60° to 180°. (middle) A set of different RNA domains, or tectoRNA units, derived from the 3D modules. Arrows indicate the structural relationship between tectoRNAs. (right) Examples of RNA shapes described in the text. The color coding for modules is the same for all panels. For more details, see also Figures S1 and S2.

Figure 2.

RNA module control of the assembly of tectoRNA units into repetitious nanostructures. (A) Assembly principle of “obtuse” or “acute” UAh_3WJ RNA units into tectosquares (TS and TSa), tectotriangles (TT and TTo), and hexameric nanorings (2TTo). Tectosquares formed of four units (A–D) assemble through four different kissing loops. Tectotriangles and hexamers are formed of three units (A, B, and T). Apparent melting temperatures (Tms) obtained by TGGE at 0.2 mM Mg(OAc)₂ are indicated for each assembly (see details on the materials and methods in the text). (B) TS, TT, and TTo constructs can assemble further through complementary 3′ tails to form programmable nanoarrays and nanogrids (Figure S3). Except for TT5 + TT6 + TT7 imaged in air, atomic force microscopy (AFM) visualization was performed under solution with 15 mM Mg(OAc)₂, as described in the text and in Figure S5. All scale bars are 50 nm. (C) Principle of assembly of AMP-TF units responsive to AMP into nanorings and fibers of different size and length. (D) Self-assembly of AMP-TF2 units in the absence or the presence of AMP at the indicated concentration is monitored by native polyacrylamide gel electrophoresis at 1 mM Mg(OAc)₂ and 10 °C. AMP-TF2 variant is shown in Figure S6. (E) AFM images of AMP-TF1 nanofibers (25 nM) acquired under solution in the presence of 2 mM ATP and 15 mM Mg(OAc)₂ at 20 °C.

Figure 3.

G1 and G2 nanohearts and their characteristic heart shapes, which are responsive to AMP. (A) G1 heart unit HY rigidifies upon the addition of 1 mM AMP. Pb(II) (42 mM) cleavage profiles of HY within G1 hearts, with (red) or without (blue) AMP, compared to those of HY alone, with (green) or without (yellow) AMP (Figure S10). Boxed 3D model of HY: regions in yellow and blue have reduced Pb(II) cleavage with AMP and units H−WXZ; regions in red retain significant cleavage. (B) AFM images of G1 hearts obtained under magnesium solution (Figure S11). (C,D) AFM images of G2 hearts acquired in air. (E) Single G2-heart images. (F) High-resolution AFM image contour-plot of a G2 heart. Each contour represents a 0.1 nm change in height. (G) The same image compared to the G2-heart 3D model. Tall points fit to the tertiary modules. (H) Cryo-EM image of G2 hearts on a thin layer of carbon. (I) Combined left and right reference-based alignments of 2456 single G2-heart particles obtained from cryo-EM images. For panels E–I, see also Figures S12 and S13. (J) G2 hearts with two different orientations of unit HA (HA1 or HA2) assemble into two distinctive dimers when linked through one tail−tail interaction. AFM images were acquired in air (Figures S8 and S14).

Figure 4.

Single-stranded G3 hearts, with structural modularity similar to G2 hearts, are produced under kinetic control in isothermal conditions during transcription. (A) Design criteria for the G3-heart KL1–6 (see the text and Figure S15). (B) Co-transcriptional folding and visualization procedure for G3 hearts. (C) Typical AFM images of G3 hearts acquired in solution according to procedure in panel B. A portion of the mica surface displayed on the right was scanned twice by AFM. Most hearts do not change much in shape (circled in blue), but a few can be disrupted by the AFM tip (circled in red). (D) Yield of well-formed G3 hearts vs malformed or partial transcripts. (E) Examples of well-resolved single G3 hearts from 0.5 μm AFM scans. (F) Examples of well-resolved single G3 hearts from 1 μm AFM scans. For panels C–F, see also Figure S16.