Genomes are complicated. Even the concept of a “gene” isn’t as straightforward as you might expect. Genes are the units of heredity, the bits of DNA and RNA that do something inside a cell. But DNA doesn’t do much of anything by itself; genes need proteins to copy themselves and to turn the small percentage of DNA that codes for a protein into enzymes. Functional parts of DNA can code for proteins or tell the cellular machinery where to start copying the chromosomes, where to start and stop the transcription of DNA into RNA and the translation of RNA into protein. In the genomes of even the simplest cells, these different components are jumbled together, overlapping with each other backwards and forwards in a dense and highly evolved sequence. While we can read DNA sequences at exponentially increasing speed and decreasing cost, understanding in detail how all these sequences are tuned to control when and where and in what conditions specific proteins are expressed is still often plodding along one PhD thesis at a time.
A paper published last week in the Proceedings of the National Academy of Science by Karsten Temme, Dehua Zhao, and Chris Voigt uses our increasing ability to synthesize DNA to tackle the problem from a very different angle. Instead of picking apart a complex set of genes to understand the regulatory details, the sequence was entirely redesigned to strip out complexity and to create a more understandable, computer-readable version of the genes. The authors started with the 23,500 base pair, 20-gene cluster that the bacteria Klebsiella oxytoca uses to fix nitrogen and “refactored” it, rewriting the sequence so that the expression of each protein-coding gene was controlled by a synthetic regulator. Refactoring is: “a term borrowed from software development whereby the code underlying a program is rewritten to achieve some goal (e.g., stability) without changing functionality.” Refactoring the nitrogen fixation gene cluster to make its regulation easier to understand and engineer involved replacing every start and stop sequence (promoter, ribosome binding site, and terminator) with synthetic versions, clustering co-regulated genes together, removing non-coding regions, and re-encoding the protein coding sequence to “create a DNA sequence as divergent as possible from the wild-type gene” and remove potential internal regulation sites.
The final refactored gene cluster is sort of a zoomed-out version of the evolved network, a lo-res cell biology textbook diagram, a pixelated genome where intricate details are abstracted into DNA chunks with defined function. Amazingly, after this extensive alteration, the refactored gene cluster was still functional, albeit at only 7.2% of the native activity. This decrease in function was expected, as the natural system has been tuned over millions of years of evolution; just because the refactored pathway is easier to engineer doesn’t mean it’s better for the cell. While the refactored system isn’t particularly good for fixing nitrogen, it does provide an interesting opportunity to study how each of the regulatory components affect the function of cluster. These kinds of experiments will likely generate data that can be fed into a computational model (now facilitated by the straightforward regulatory design) that can enable both better reading and writing of genetic systems.
Refactoring has played a role in synthetic biology for a few years, but I find this story especially interesting now in light of the recent discussion of the New Aesthetic, which describes art and design that celebrates the “eruption of the digital into the physical.” Synthetic biology re-imagines and refactors living systems to be more like computers, with DNA functioning as code used to program living hardware according to logic diagrams and designed with the help of rigorous computational models and CAD software. Synthetic biology is the eruption of the digital into the living.
The New Aesthetic is an interesting lens through which to understand the way that synthetic biology translates living systems into computer models that are then retranslated back into engineered cells. This flip-flop between the living and digital world can generate biological insight and useful biotechnologies, but it also leaves behind interesting artifacts of the transition. The DNA code of the nitrogen fixation gene cluster was sequenced, the sequence was edited and modeled in a computer, redesigned according to engineering principles and put back into the cell, functioning in a clearer but less robust way. We gain control but we lose function.
Biological pattern formation makes it easier to visualize these bio-digital transitions. Biological logic governs the patterning of cells, for example the location of spots on a cheetah relative to one another. These rules can be translated into computer models that generate Turing patterns, patterning pixels rather than cells. These models in turn can inspire the design of synthetic gene networks in bacteria that control the simple patterning of red and green fluorescence in a petri dish—leopard print bacteria.
These transitions can also influence other kinds of design inspired by synthetic biology. Synthetic Aesthetics brings together artists and scientists to explore collaborations between synthetic biology, art and design. Residents Fernan Federici, a plant scientist at Cambridge University, and David Benjamin, an architect at Columbia University, developed a fascinating project that turns a computational model of a living system into a physical object. They generated models of the growth of xylem cells, specialized structures that transport water throughout plants. These models then were applied to the shape of buildings, constraining the way that the structure could grow to fill the space. These computational structures were then 3D printed, bringing the bio-digital into the physical world.
I see New Synthetic Aesthetics (and Ethics) playing a role in how the lo-res synthetic systems we design today might transition into more fully rendered structures and behaviors—how we will understand, model, design, and engage with these new creatures. Adding a detour between the digital and the physical in the biological world may help make the New Aesthetic if not weirder, then certainly squishier and smellier. With the New Aesthetic we’re seeing the cracks in the barrier between the digital and the physical; merging this with the biological in all its complexity will generate a new set of cracks, new assemblages and a new set of artifacts, from Turing pattern bacteria to E. coli that play Sudoku or solve the Hamiltonian Path Problem. Maybe I’ll have to start a new Tumblr…