With three-dimensional structures such as this one, the real challenge isn't designing the object but proving that it successfully formed, says Paul Rothemund, a computer scientist at the California Institute of Technology, who developed a simple technique for making DNA structures. The researchers used several different imaging methods to ensure that the boxes assembled themselves as planned. "They did a very convincing job of showing that they made what they thought they made, which is really important," Rothemund says. "And now they're free to try and elaborate on it and get it to actually do something."

Kjems has several ideas for what the boxes might do. One possibility is to load them with drugs and program the lids to open in response to some biological cue inside the body--the presence of a virus or a cancer gene, for example--thereby releasing their therapeutic cargo.

"There's a way in which they're more interesting than almost any other nano-encapsulation scheme you can think of for that purpose, because they have these infinitely programmable lids," says Rothemund. "That's something that no other kind of nano-drug-delivery capsule can offer."

Therapeutic uses are still a long way off, however. While the boxes are theoretically solid enough to prevent large molecules from leaking out and spacious enough to enclose a ribosome or a small virus, the researchers haven't yet tried to put anything inside them. And so far, the boxes only function inside a test tube. Unlike some other nano-delivery vehicles, there's no evidence yet of the safety or efficacy of DNA-based devices in living systems.

But the lockboxes needn't carry a payload to prove useful. Kjems also envisions turning them into electronic components. Because they have two distinct keys, the boxes act as AND gates, opening (and glowing green) only when both keys are present. With a few straightforward tweaks, they could serve as NOT gates or OR gates as well. "In principle," says Kjems, "you could build a DNA computer using these boxes."