A complex DNA nanostructure is amplified and folded in bacterial cells. Image courtesy of H. Yan.

“Let the cell do the work” was the mantra behind recent efforts to express complex DNA nanostructures in bacterial cells, led by Hao Yan of Arizona State University and Nadrian Seeman of New York University.

Two features of DNA make it an ideal molecule for the assembly of complex structures: its code of only four nucleotides is easy to program, and the resulting sequence is structurally predictable and should yield two-, even three-dimensional constructs for many different applications. The current bottleneck to actually achieve a higher-level production of DNA nanostructures is a limit to the scale at which long single-stranded DNA molecules of complex secondary structure can be produced.

Yan explains why it is important to scale-up production: “We want to build conformational information into the DNA nanostructure and allow the DNA to self-assemble into an algorithmic pattern.” Sensing that this rather technical outlook may fail to excite outsiders to the field, he adds: “Look at the growth of a tree, it stops at the tip so it must have an algorithm of growth built in. DNA can potentially do the same thing; the big question is how to scale-up self-assembly to larger domains that contain more information.”

The current state of the art is still very far removed from this 'DNA tree' image. Yan and his colleagues have assembled DNA nanostructures in vitro but could not easily exceed a length of ∼120 bases of single-stranded DNA. To get longer pieces at higher yield, the researchers shifted the workload from the test tube to a bacterial cell. Yan's first question was, “Will nature's molecular machinery tolerate the complex structure?”

The answer was an unequivocal yes. Not only did E. coli produce long DNA strands in large quantities, it also tolerated their assembly inside the cell. Yan and his colleagues demonstrated that the clover-like structure programmed into one DNA molecule was present inside the living cell.

Excited by this finding, Yan wants to fully exploit the possibilities of an in vivo production system and apply selective pressure so that the nanostructure can evolve. His end goal is a functional, not a static, structure with a myriad of applications, ranging from single-cell imaging, where a tweezers-like nanostructure can bring together two fluorescent proteins for energy transfer, to protein detection with a flexible structure containing an induced fit for a certain protein, to an aptamer that would recognize and neutralize a bacterial or viral intruder in the cell.

These intriguing applications in mind—some of which are, as Yan acknowledges, still closer to fiction than to science—the researchers now seek to lay the necessary groundwork. For example, longer DNA molecules result in an increased error rate, consequently proofreading mechanisms are needed. The vector and promoters used for the in vivo replication must be carefully chosen; Yan cautions that with larger molecules one must make sure that the sequence of the vector does not interfere with the sequence of the nanostructure. So far the team has only tested the replication of nanostructures in E. coli; it is still an open question whether eukaryotic cells will support production and assembly.

The field of DNA nanotechnology, comprised of about ten research groups, is relatively small, but their dreams, and the implications for society, certainly are not.