Tuesday, August 19, 2025

RNA Origami

Nature’s RNA Origami: Cryo-EM Reveals Giant, Protein-Free Machines Hidden in Bacteria

Image by TheKuriousK

For decades, we’ve assumed that large, well-ordered macromolecular machines are the domain of proteins (with a few viral RNA outliers). This paper blows that assumption open. Using high-resolution cryo-EM, the authors show that three huge bacterial noncoding RNAs—OLE, ROOL, and GOLLD—self-assemble into symmetric, protein-free megastructures:

  • OLE forms a dimer—two long, coaxial “pipe bundles” welded by unusual RNA-RNA bridges.

  • ROOL self-organizes into an octameric nanocage (≈28 nm wide), entirely empty inside except for a flexible internal loop.

  • GOLLD builds an even larger tetradecameric nanocage (≈38 nm), again RNA-only.

The team also shows these oligomeric stoichiometries persist at very low concentrations, and comparative genomics points to evolutionary conservation of the intermolecular contacts that make the cages and dimer possible. Translation: these are probably not in-vitro curiosities; they likely matter in cells. 

Why this is a big deal

Most of what we know about RNA structure comes from small motifs, riboswitches, ribozymes, ribosomes, and engineered scaffolds. Despite >4,000 RNA families in Rfam, only a sliver have had their tertiary structures solved, and even fewer have quaternary architectures mapped. This work expands that map dramatically—and, crucially, it shows that multimeric RNA-only assemblies can be big, symmetric, and functionally conserved in nature, not just in synthetic nanotechnology demos. 

It also lands amid a burst of convergent evidence: an independent Nature paper this summer reported ROOL as an octameric natural RNA nanocage, with dimensions and internal disorder that echo the present study. Another contemporaneous report analyzed higher-order natural RNA-only multimers more broadly. The field is clearly coalescing around a new idea: RNA can natively build machines—without proteins—that rival protein complexes in size and symmetry.

Meet the three “ornate” RNAs

1) OLE: the pipe-bundle dimer that scaffolds partners

OLE (Ornate, Large, Extremophilic) RNA has been a biological puzzle for years, implicated in stress responses in alkaliphilic and halophilic Gram-positives and known to localize to membranes and bind multiple proteins (OapA, OapC, bS21/RpsU among others). The mechanistic picture was blurry—until now. 

In this study, a 577-nt OLE from Clostridium acetobutylicum produces compact cryo-EM particles whose 3D reconstruction at 2.9 Å reveals a C2-symmetric dimer. Each monomer contributes a bundle of long, parallel A-form helices; the two “pipes” are welded by three striking RNA–RNA bridges:

  • B1: an unusual, conserved A–A base-paired helix,

  • B2: stacked loop–loop interactions, and

  • B3: a kissing-loop between identical hairpins.

Image source: Kretsch, et al.  Nature 643 CCBY  

Together these bridges lock the two halves into a rigid scaffold. The structure also exposes pockets where known OLE-binding proteins could dock—without the RNA needing to refold—reframing OLE as a pre-organized RNA organizer for its protein partners rather than an RNA that requires proteins to fold. The authors even superpose membrane protein OapA and accessory OapC in plausible positions, hinting at a higher-order RNA–protein assembly at the membrane. 

Why it matters: OLE’s architecture suggests a modular RNP hub: an RNA dimer forms the backbone; proteins clamp onto pre-formed RNA motifs to create a specialized membrane-associated machine, consistent with earlier genetic and localization data that tied OLE to stress adaptation and alcohol tolerance in extremophiles. 

2) ROOL: nature’s octameric RNA nanocage

ROOL (Rumen-Originating, Ornate, Large) RNAs live in bacterial and phage genomes, often near tRNA islands. Their predicted secondary structures have long looked “too ornate to ignore”—dense with pseudoknots and conserved loops—but no protein partner was known, leading to speculation that ROOL might be RNA-only. The new cryo-EM reconstructions confirm that: ROOL forms an 8-mer (D4 symmetry) nanocage, roughly 28 nm in diameter, larger than the bacterial ribosome’s longest dimension. Each 659-nt monomer is relatively flat, but strings itself to neighbors through a lattice of quaternary interactions: multiple kissing loops, A-minor contacts, and a daisy-chain of stem-loops that forms an inner ring at the top of each half-shell. A disordered internal linker dangles into the cage’s hollow interior. 

Particularly interesting are the inter-half-shell bridges: each monomer reaches across to two monomers in the opposite hemisphere, biasing assembly toward the full cage rather than isolated halves or dimers—a clever architectural constraint nature often uses in protein capsids. An independent group reported a near-identical ROOL octamer earlier this summer, supporting the generality and robustness of the structure. 

Why it matters: ROOL’s cage suggests a microcompartment function—RNA forming a shell that sequesters or delivers molecules. The interior is big enough for metabolites, tRNA arrays, or even parts of ribosomes; its flexible linker hints at cargo capture or gating. Think “RNA carapace,” not just regulatory hairpin. 

3) GOLLD: the even bigger 14-mer cage

GOLLD (Giant, Ornate, Lake- and Lactobacillales-Derived) RNAs are the largest of the trio (often >800 nt), also phage-associated and frequently co-localized with tRNA cassettes. Cryo-EM reveals a tetradecamer (14-mer) with D7 symmetry and a whopping ~38 nm diameter. Like ROOL, GOLLD is a hollow cage with a single disordered loop in the interior; unlike ROOL, its 5′ and 3′ domains segregate into distinct architectural zones within each half-shell (cap vs. ring). 

The quaternary glue repeats the familiar RNA toolkit—kissing loops and A-minor insertions—but the wiring diagram differs. For example, a conserved cap interaction (B2) creates a daisylike inner ring (with a shorter spacing than ROOL’s to accommodate 7 monomers per half), while a set of three inter-hemisphere bridges (two self-kissing loops and one A-minor contact) cinch the two bowls into a sphere. The end result: a ribosome-sized, all-RNA cage with evolutionarily tuned geometry.

Why it matters: the domain separation between GOLLD’s 5′ and 3′ regions may explain sequence divergence patterns seen by comparative genomics; the domains can evolve semi-independently as long as inter-domain and inter-chain contacts are preserved. That’s the signature of functionally constrained quaternary design. 

Are these structures biologically real—or artifacts?

The authors devote considerable effort to this question and offer five lines of evidence:

  1. Controls with other RNAs. Under the same cryo-EM conditions, a different large bacterial RNA (the raiA motif) reconstructs as a monomer, arguing against a generic “everything oligomerizes on the grid” artifact. Similarly, HEARO RNA lacks structure without its protein partner, reinforcing that not all large RNAs collapse into multimers in vitro. 

  2. Mass photometry at nanomolar concentrations. OLE, ROOL, and GOLLD show discrete mass peaks corresponding to 2-, 8-, and 14-mers down to ~12.5–50 nM, far below the concentrations used for grid prep and well below typical cellular copy numbers for these RNAs—meaning the oligomeric states are favored even when molecules are scarce. 

  3. Dynamic light scattering (DLS). For ROOL and GOLLD, DLS finds thermostable, monodisperse particles with no detectable monomer fraction up to 55 °C, consistent with stable cages rather than fragile aggregates. 

  4. Conserved, multivalent interfaces. Each monomer contributes 5+ distinct inter-subunit contacts; these are sequence- and structure-conserved across homologs (covariation supports the precise base pairs within kissing loops and A-minor adenosines). Conservation at intermolecular positions is the smoking gun for selected quaternary architecture.

  5. Independent replication by others. The ROOL octamer was solved independently at similar resolution, with matching dimensions and cage topology, strengthening the biological case. 

Bottom line: these assemblies aren’t artifacts—they look like real, evolved machines.

The RNA “quaternary code”: motifs that build megastructures

Across OLE, ROOL, and GOLLD, the same limited vocabulary of RNA tertiary interactions gets reused in modular ways:

  • Kissing loops: cognate hairpins that base-pair to dock chains.

  • A-minor interactions: adenosines that insert into minor grooves of neighboring helices, acting as corner pins.

  • Non-canonical stacks and A–A helices: local “welds” that stiffen interfaces or bridge symmetry mates.

It’s the geometry of where these motifs are placed—distances between repeat elements, loop lengths, and helical orientations—that determines stoichiometry and symmetry (C2 vs. D4 vs. D7) and whether you get dimers or cages. This is analogous to protein quaternary design rules (helix-helix packing angles, interface repeats), but executed with RNA’s own toolkit. 

What could these RNA machines do?

The function of all three classes remains enigmatic, but the structures inspire plausible hypotheses:

  • OLE as an RNP hub at membranes. The dimer exposes kink-turns and pockets that match suspected binding sites for OapC, OapA, and bS21/RpsU, suggesting OLE organizes a membrane-associated stress-response complex. Superposed models even admit a scenario where OLE scaffolds higher-order assemblies of OapA, faintly reminiscent of double-stranded RNA channels (think SID-1-like topologies). That dovetails with older genetics linking OLE to alcohol tolerance and stress adaptation. 

  • ROOL & GOLLD as RNA microcompartments. Both cages are empty except for internal flexible linkers that project into the lumen—prime real estate for binding cargo. The diameter is too small for a full phage genome, but ample for tRNA arrays (sometimes genetically fused to GOLLD), small RNA–protein complexes, metabolites, or even transient ribosomal pieces. The authors explicitly propose encapsulation and opening/closing of the two half-shells (few inter-hemisphere contacts = easy hinge) as a functional cycle. 

  • Information economy. The strong covariation signals at inter-chain contacts imply that cells (or phages) are maintaining the cages across evolution—suggesting fitness value, not accident. That value might be cargo sequestration (e.g., protect tRNAs from nucleases during phage infection), local concentration of factors, or buffering of stress metabolites. 

Encouragingly, the independent ROOL study reported very similar cage dimensions and observed poorly ordered internal regions, also consistent with cargo-friendly interiors. 

Methods in a nutshell: how do you solve RNA megastructures?

Achieving sub-3 Å for big RNAs is hard. The team combined:

  • In-vitro transcription and Mg²⁺-assisted folding to produce monodisperse particles.

  • State-of-the-art cryo-EM (Titan Krios + energy filters; Falcon 4 or K3 cameras), followed by non-uniform refinement and symmetry expansion in cryoSPARC to improve per-asymmetric-unit signal.

  • RNA-aware modeling (ModelAngelo seed + Coot/ISOLDE/ERRASER2 refinements) to thread sequences and resolve non-canonical pairs and motifs.

  • Biophysics (mass photometry, DLS) to confirm stoichiometries at nanomolar levels and thermal stability.

  • Comparative genomics and covariation (Infernal, R-scape, RNAalifold) to show that intermolecular contacts are under selection.

The independent ROOL paper followed a similar cryo-EM/refinement path and deposited maps and models consistent with an octamer. 

Implications beyond this paper

  1. Rewriting the RNA design playbook. If nature routinely uses kissing loops + A-minor pins to enforce D-symmetries at 20–40 nm scales, RNA nanotechnologists have a native template for building drug-delivery cages, encapsulation shells, and logic-gated containers—with sequence-programmable openings. The new data give priceless atomic geometries for stable interfaces. 

  2. Origin-of-life & early evolution. Protein-free, symmetric RNA megastructures capable of encapsulation resonate with hypotheses where RNA handled compartmentalization and molecular crowding before proteins took over. Cages could have concentrated cofactors or protected labile intermediates. While speculative, the plausibility just went up. 

  3. Cell biology of bacteria and phages. If ROOL/GOLLD are phage-encoded, an RNA cage could be a rapidly assembled, translation-free microtool the phage deploys in the host: e.g., to corral host tRNAs, sponge regulators, or stage ribosomal parts. Conversely, bacterial OLE might be a stress-gated docking platform that reconfigures membrane proteins on the fly. The independent ROOL study and related follow-ups will likely fuel in vivo tests soon. 

  4. Computational structure prediction. These high-quality maps and atomistic models supply much-needed training/benchmark data for RNA 3D prediction—especially for quaternary modeling and symmetry inference, long-standing blind spots compared to proteins. 

What we still don’t know (and how to find out)

  • What’s inside the cages in vivo? Crosslink-mass spec, in-cell SHAPE/PAINT, proximity labeling from within the lumen, and native immuno-purification could identify cargo.

  • Do the cages open? Single-particle cryo-EM of partially assembled states, time-resolved EM, and AFM on supported membranes might capture hinge motions.

  • Are there gates or triggers? Mutate kissing loops and A-minor pins to test for pH/ion-dependent opening; screen for ligand-induced conformational changes.

  • What’s OLE doing at membranes? Reconstitute OLE with OapA/OapC in nanodiscs or proteoliposomes; test whether OLE scaffolding modulates conductance or lipid curvature.

  • How widespread is this? Mining metagenomes with covariance models tuned for quaternary motifs may reveal a zoo of cages and dimers hiding in plain sight. 

A quick guide to the key terms

  • Kissing loop: Two hairpin loops that base-pair via their exposed nucleotides, docking the helices.

  • A-minor interaction: An adenine inserts into the minor groove of a helix, stabilizing tertiary contacts.

  • D-symmetry: Dihedral symmetry; think “two opposed N-fold rotational axes” (e.g., D4 = 8-mer cage).

  • Linker/disordered loop: Flexible segment often used to bridge distant motifs or project into cavities.

  • Quaternary structure: How multiple RNA chains assemble into a larger complex.

The bigger picture: RNA as an architect, not just a message

The classic image of RNA is “information and regulation”: mRNAs, tRNAs, ribozymes, riboswitches, miRNAs. This work, together with the independent ROOL study and a broader survey of natural RNA-only multimers, adds a complementary identity: RNA as a structural architect. Not merely a scaffold that waits for proteins, but a polymer that can plan, place, and pin its own mega-assemblies—and do so with a small motif vocabulary used in precisely repeated patterns. 

If you are designing RNA nanodevices, you’ve just been handed a nature-validated parts catalog: symmetric kissing-loop rings, A-minor cornerstones, and spoked caps that set cage diameter and stoichiometry. If you’re a microbiologist, you’ve got new hypotheses for stress-response architecture (OLE) and phage host-takeover tools (ROOL/GOLLD). And if you’re a structural biologist, you’ve got a front-row seat to the next wave of RNA machines.

Further reading

  • Kretsch et al. (2025) Nature — the study summarized here; includes beautiful maps, models, and exhaustive methods.

  • Ling et al. (2025) Nature — independent ROOL octamer RNA nanocage with matching dimensions; a great companion read. 

  • NSMB article (Aug 2025) — survey of higher-order natural RNA-only multimers, situating these discoveries within a broader landscape. 

  • Background on OLE biology — OLE in extremophiles, membrane localization, and stress resilience; contextualizes the OLE dimer as a membrane-adjacent organizer. 

Key takeaways

  1. RNA can be its own engineer. It doesn’t need proteins to build big, symmetric machines.

  2. Three classes, three architectures. OLE dimer (C2), ROOL octamer (D4 cage), GOLLD tetradecamer (D7 cage).

  3. Interfaces are conserved. Kissing loops and A-minor contacts form a reusable quaternary code.

  4. Likely biological roles. OLE as membrane RNP hub; ROOL/GOLLD as encapsulation devices.

  5. New frontier. These structures will reboot RNA nanotechnology and refocus in vivo studies on RNA-built microcompartments. 

Study: Kretsch, R.C., Wu, Y., Shabalina, S.A. et al. Naturally ornate RNA-only complexes revealed by cryo-EM. Nature 643, 1135–1142 (2025). https://doi.org/10.1038/s41586-025-09073-0


Keywords: RNA origami, RNA nanotechnology, RNA self-assembly, RNA-only complex, natural RNA multimer, RNA nanocage, OLE RNA, ROOL RNA, GOLLD RNA, cryo-EM RNA structure, A-minor interaction, kissing loop, A–A helix, pseudoknot, ribose zipper, dihedral symmetry, D4 symmetry, D7 symmetry, C2 symmetry, quaternary RNA structure, noncoding RNA, bacterial RNA, phage-encoded RNA, microcompartment, membrane-associated RNA, OapA, OapC, RpsU, mass photometry, dynamic light scattering, covariation analysis, Rfam, comparative genomics, extremophiles, tRNA islands, stress response, membrane scaffold, RNA design, RNA cage

Hashtags: #RNAOrigami, #RNANanotechnology, #RNA, #ncRNA, #RNAstructure, #CryoEM, #StructuralBiology, #MolecularBiology, #SelfAssembly, #Nanocage, #QuaternaryStructure, #KissingLoops, #AMinor, #Pseudoknot, #RiboseZipper, #OLE, #ROOL, #GOLLD, #PhageBiology, #Metagenomics, #ComparativeGenomics, #Bioinformatics, #MassPhotometry, #DLS, #Rfam, #Extremophiles, #RNAEngineering, #SyntheticBiology, #RNADesign, #MembraneProteins, #OapA, #OapC, #RpsU, #Bacteriophage, #StressResponse, #Nanobiology

at
Labels: , , , , , , , , , , , , , ,

No comments:

Post a Comment

Subscribe to: Post Comments (Atom)

Featured Story

Viral Mystery: The Case of the Missing Molecule