As my recent Technology Feature made clear, the technology to synthesize large genomes is advancing at a remarkable pace. So too are technologies for wiring genetic circuits to endow those genomes with novel properties. In the 3 August issue of Nature, researchers at Arizona State University in Tempe describe a new technology to do just that.
Synthetic biologist Alexander Green of the Biodesign Institute at ASU, and colleagues, describe simple ‘ribocomputing devices‘ that can function as logical AND, OR, and NOT gates — circuit building blocks that control the translation of a reporter gene based on the presence one or more small input RNAs. As those inputs can reflect exposure to different environmental agents or stimuli, the system could serve as a kind of biological sensor.
Synthetic gene circuits typically are built up from smaller modules using transcription factor intermediaries or recombinases. But according to Green, RNA offers several advantages. It folds in predictable ways, is single-stranded (and thus available for complementary base-pairing), and “exists in the middle of the Central Dogma,” being written in DNA but coding for protein.
The team designed their ribocomputing devices based on ‘toehold switches‘, a technology they first described in 2014, in which segments of RNA provide landing pads (‘toeholds’) for short complementary sequences. Binding of those complementary molecules (called ‘trigger RNAs’) unwinds a stem-loop structure near the 5′ end of a longer mRNA (the ‘gate RNA’), thus exposing a ribosome-binding site and activating translation of a downstream reporter gene (in this case, the fluorescent protein, GFP). By combining two or more such landing pads, researchers can build ever more complicated circuits.
In their simplest design, the team placed two landing pads upstream of GFP such that binding of either trigger RNA was sufficient activate the reporter, creating a logical OR device. For the AND gate, two binding events are required to activate the reporter, while the NOT configuration employs two trigger molecules that are complementary to one another, such that the presence of one input blocks the activity of the other.
The team then moved on to more sophisticated designs. The most complex was capable of evaluating the expression (A1 AND A2 AND NOT A1*) OR (B1 AND B2 AND NOT B2*) OR (C1 AND C2) OR (D1 AND D2) OR (E1 AND E2) — a 12-molecule problem for which they tested 28 different combinations of inputs. Yet for all that complexity, the system appended just 400 nucleotides to the 5′ end of the GFP transcript, Green says. “That’s one of the nice things: It’s actually quite compact.”
The team took a structural approach to design its RNAs, first deciding on a desired shape and then working out the sequence to create it. An online tool called Nupack provided the computational muscle to design the actual sequences. “RNA offers this higher degree of programmability,” Green says. “So we can use computer-based design and really get very close to what we theoretically expect its behavior to be in an actual experiment.”
Though the paper was a proof-of-principle using wholly synthetic triggers, future designs could act instead as sensors for endogenous sequences such as viral transcripts or cellular RNAs that indicate exposure to a particular environmental condition, Green says. Such designs could even be combined with transcription factor and recombinase-based designs to build more sophisticated circuits, he adds.
“We’re showing how far the system can go,” he says. “And now that we’ve done this, the things that are a little more realistic are actually simpler. We know that our system is capable of handling them.”
Jeffrey Perkel is Nature‘s Technology Editor.