A DNA origami box
Design (A), with a cut-open and
full view (B, C) of a cryo-EM map*
DNA box with a controllable lid”
ES Andersen, et al., Nature, 459:73–76 (2009).
A paper in the current issue of Nature reports the fabrication of DNA origami boxes ~35 nm on a side; to my knowledge, these are the first closed structures made by these means.
Paul Rothemund’s initial 2006 Nature article [pdf] described a methodical way to form patterns of raised features on origami, based on a raster pattern with ~200 distinct addressable pixels per tile. The attachment points in this technique are the ends of the staple strands that bind one helix to the next, and they are therefore individually addressable — In principle, each raised feature could have a unique structure. In Paul’s work, the features were folded DNA, extensions of the strands themselves, but this was a convenience and by no means a constraint. Origami-based instances of modular molecular composite nanosystems could employ this to position a wide range of functional components.
Structures like the box in the present work are built of the same motifs and will afford the same capability, but provide attachment points that are no longer constrained to a plane surface. Modular molecular composite nanosystems built on this basis can therefore be compact and three dimensional.
The present work also illustrates one way in which origami structures can be reconfigured after synthesis: Boxes can be configured to open by means of externally supplied DNA ‘keys’.
Of greater interest from a fabrication perspective, however, is the inverse behavior, fastening origami components together by means of DNA strands added after an initial assembly process. This can enable the fabrication of compact self-assembled systems in which premature completion of a confining structure would interfere with subsequent assembly steps. Delayed (or sequenced) fastening provides a means to design assembly processes, not just their products, and thereby avoid a range of potential problems. The authors demonstrate this this behavior, too. They assemble the sides, linked only by their corners, before adding the staples that stitch together the seams to complete the box.
* Single-particle cryogenic electron microscopy (reviewed here [pdf]) is an alternative to cryo-electron tomography, which I discussed in an earlier post.
{ 2 comments… read them below or add one }
Seems to open up tremendous opportunities. Im just wondering if this concept could lead to a simple factory concept. Simply stitch the lock boxes together in a massively parallel grid, itself forming the base of a larger lock box configuration, itself part of another massive grid of such boxes, and continue layering up. get the software to program the constructs so that at each level they are built with the next levels staples ready to assemble the higher level configuration. The actual base boxes might be better with 2 doors, one for inserting the starting building materials, doesnt need to be these lockboxes, could be tubes, with some smarter filter setups to automatically intake the right dna pieces which the design software program has designated is required (im only using lock boxes to keep the whole thing based upon something that has been done, but of course a much better designed setup could be constructed/designed).
Anyhow to continue, theres no need for the boxes in each layer to produce the same components, they could assemble any mix of components for assembly at the next layer. Assuming good designs can be made to ensure things are robust, and the right mediums are used, is there even any reason to prevent this scaling to almost macroscales. I could imagine the last size Lockbox sitting on the layers might be a speck sized square millimetre, with its components floating in solution and assembling after a little shaking. Maybe with the complexity that such components might be able to hold, there may not be so much need to go to macroscale fabbing anyway, as soon as you are able to assemble microscale structures that can have effective motion/motors controlled by some CPU equivalent (who knows maybe even lockbox based if as the authors point out could process info), then you could create claytronic or fractal robotics which can assemble themselves into larger assemblies. Best of all the wonderful concept of foglets. I’d love to hear you explain why this is just all wrong, because currently, though it sounds outlandish, I dont really see it as being beyond the bounds of technical possibilities, extending from what the authors have said. All the best, Mark
Mark — There’s an enormous mismatch between available technologies and systems at anything approaching that level of scale and complexity. The agenda today is to develop a far more basic engineering methodology that builds directly on available or one-step-removed laboratory capabilities.
{ 5 trackbacks }