At the recent Foresight conference, Seeman showed a DNA machine that would twist back and forth depending on whether a strand was bound to it. And the other side of the machine would catalyze strands to link together. (I forget whether there was another enzyme involved in the linkage, or whether it was just driven by the huge effective concentrations you get with positional control.) Depending on whether the machine was twisted back or forth, two different sequences could be made.

So this machine still delivers less than one bit-per-building-block, but not by much! And although at first sight it seems like a toy proof-of-concept, I can think of at least two ways to make it more interesting. One: present tRNA binding sites, and use it to make proteins. Two: instead of just two selectable sequence binders, make a rotor with lots of them.

Chris

More information on the RNA tectonics & JigSaw Puzzle self-assembly can be found at physorg.com here.

It looks like the day is fast approaching when we could manufacture ~2 kg of the Liao/Seeman assemblers (or derivatives thereof) and send 1 kg to both Smalley & Whitesides as Christmas presents….