Second Friday afternoon panel on Applications of DNA Computing and Molecular Programming
Beckman Institute Auditorium, 2:15pm to 3:00pm, Friday.
Back to Main_Page
The entire field has the general problem that no one thinks it's useful for anything. There, I said it. And it's sort of true, at least for now. Seeman's / Adleman's visions (now multiplied by the many) have yet to become reality. There are several reasons for this problem: (a) DNA computation in and of itself isn't very useful, compared to silicon computation. The whole "many molecules working in parallel" thing didn't get very far. I'd be happy to take on Erik and Lulu with my hand calculator any day of the week; (b) DNA is not exactly the world's best building material. The notion that you're going to make silicon-like devices with DNA may yet work out (go Rothemund and others!), and DNA costs continue to come down, Moore's law, yada, yada. Still. I just don't see making a programmable support beam for a skyscraper with DNA. I don't necessarily even see making a good photoresist with DNA, although that's more likely.
Well, that's a bit of a bummer. How do we go forward?
My own answer is diagnostics for resource-poor settings. I think that the amplifiers that Winfree, Pierce, Yin, and others have built are very cool. No enzymes, isothermal, programmable, low cost of goods ... spit in a tube and readout the result. We're working hard on this, and so are others. That said, I do not see DNA circuitry being competitive against standard technologies, such as PCR, in settings where you have a cold train. Enzymes are very good at what they do, which brings up the notion that we really need to get enzymes much more involved in DNA computation again.
But I think there is a bigger, better answer waiting in the wings, and that involves new materials. DNA is unique. Weak and expensive and a pain-in-the-butt, but unique. There will be DNA-like molecules into the future where the lessons we learn in programming today will pay off. I actually do believe we'll have programmable support beams (or somesuch), but the way in which we incorporate programmability into materials science has yet to be built.
When I talk to others, they have good answers, too, mainly centered around that other amorphous field, nanotechnology. DNA tools for nanotechnology is a bit like selling plastic shovels at a fool's gold rush, but there are indications that DNA probes and tips and actuators and whatnot are going to be making real contributions to science in the near future.
DNA Computing and Molecular Programming applied to Living Systems.
Computers with silicon chips look very different from the cell. Yet the cell is also a computer – a cell and indeed groups of cells can take a decision based on a set of inputs. Every action of a cell, for example, the upregulation or downregulation of a protein in response to a molecular trigger may be viewed as a computation. The specific molecular triggers may be considered as inputs and the resultant molecular change i.e., protein upregulation or downregulation would be the output.
A plethora of synthetic DNA-based devices that operate on molecular logic could be envisaged to report on or manipulate functional states of the cell by detecting or interacting with functional RNA molecules. However, one of the major obstacles in translating this is in the real world, is the interfacing DNA-devices with biological systems. For this one is limited by the precise and easy delivery of DNA architectures within the biological systems.
Interestingly, many of the principles of structural DNA nanotechnology are increasingly being co-opted into the RNA domain to yield exciting RNA-based devices. The huge advantage of working in RNA space is the ability to genetically encode these devices thus doing away with the problems of DNA-device delivery. I welcome views on my opinion that the integration of the powerful principles of DNA computing and programming into RNA space might be a way forwards to both manipulate and understand cell function.
It's too late to question whether molecular programming will ever be good for anything. It already is. But more importantly, if we collectively pick our battles and focus on applications where conventional approaches fall short and programmable molecules offer new hope, this decade will end not with scattered applications examples, but with the beginnings of a technological revolution identifiable for all to see. This is not to say it's an easy road ahead. There's a big difference between a proof-of-principle demonstration in expert hands and a programmable molecular technology that robustly addresses the needs of non-experts. Traversing this divide leads to unfamiliar multi-disciplinary challenges, but creativity, unconventional collaborations, dogged pursuit, and time, provide a hair-raising recipe sufficient to overcome prevailing skepticism.
Technical challenges toward lab on a molecule
It has become appealing to implement "lab on a molecule" by combining structural DNA nanotechnology with DNA computing. When molecules/biomolecules are arranged to interact with each other with spatial control (~6 nm resolution for DNA origami and ~1-2 nm if we are talking about the end of neighboring helices), the well structured giant molecule can sense, compute and actuate. While we continue to search other types of information coding molecules that could offer angstrum resolution, there are several technical challenges we need to overcome to use the existing structures to achieve the "lab on a molecule". First, we are limited by methods to make reliable site-specific protein-DNA conjugates. Most current strategies work only for a few types of proteins. One often needs to develop a new protocol if he/she is interested in using a different protein. This limits our ability to attach interesting protein molecules to DNA nanostructures and study their interactions. Further, even one could conjugate DNA oligo reliable to a specific site on the protein surface, it is still difficult to control the orientation of the protein molecule displayed on a DNA nanostructure surface. The same challenge applies to nanoparticle (nanowire, nanoprisms)-DNA conjugation. Surface chemistry can be quite messy, all sorts of problems can happen (aggregation, ligand exchange, to name a few), reliable methods to make robust linkage between DNA with variety of inorganic materials are needed. Recent progress in in-situ synthesis of quantum dots with controlled number of DNA oligos on the surface is promising, but more work needs to be done in this area to allow us to construct multifunctional and multicomponent DNA templated structures with high yield and precision.
While we are dealing with these challenge, structural DNA nanotechnology has already began to demonstrates its promise in various applications. DNA nanostructures can be used for molecular and cellular biophysics (e.g. Seeman's DNA scissor can measure the protein/DNA binding energy, Sugiyama used DNA origami frames as molecular ruler to measure the dynamic process of enzyme binding), for creating bio-mimetic system (e.g. engineered nanopore to connect cells); for energy transfer and photonics (e.g. light harvesting complexes), for diagnostic and therapeutics (e.g. hybridization chain reaction based detection and controlled drug release using DNA cages).
Expression and Assembly in vivo
Of great relevance to many cell biologists would be the biocompatibility of DNA nanostructures and their potential for function in cells. Naturally this raises the question of whether such nanostructures can be genetically encoded for intracellular expression and assembly. Thus far, it has been shown that DNA nanostructures encoded as long single strands, designed by taking advantage of the paranemic crossover motif, can be amplified by polymerases in vitro or in vivo. It remains to be shown that increasingly complex DNA nanostructures can be folded efficiently within a cellular context. The emerging field of RNA nanotechnology might seem more promising in this regard since RNA is readily transcribed into a single strand, which can be directly folded into a programmed nanostructure; DNA can be made in a single-stranded form as well, although less commonly, using rolling circle or else reverse transcription based methods. RNA nanotechnology is still in its infancy. Even though the rules for the reliable design of RNA nanostructures are comparatively scarce, a recent work by Silver and co-workers represents a significant step in engineering RNA molecules to assemble into pre-defined discrete 1D and 2D structures in vivo. More importantly, DNA and RNA nanotechnology have enormous synergistic potential, where the predictability of DNA folding can be coupled to the diversity of RNA functionality. The next challenge is to combine the structural assembly with molecular computing for the in vivo assembled system.
DNA17 participants, please enter comments and questions here.