Teruo Fujii and Yannick Rondelez at the University of Tokyo, Japan, have been working on chemical oscillators that can create a simple circuit – the product of a reaction feeding back and affecting that one or subsequent reactions. Rondelez explains that many physical situations, from chemical dynamics to ecosystems, can be described by similar mathematics. ‘While the analogy between chemical and biological systems does not need a huge mental stretch,’ he says, ‘I thought it would be nice to put the finger on the one between chemical networks and ecological ones.’
The ecological system Fujii and Rondelez modelled is a simple predator–prey one. As the number of prey in the ecosystem grows, the number of predators also grows. This which then leads to a drop in the prey as their numbers dwindle from hunting, followed by a decline in predators as they 'starve'. To model this using molecules, Fujii and Rondelez turned to small lengths of DNA. The advantage of DNA, says Rondelez, is that you can tailor the DNA to build circuits that wouldn’t be possible using other small molecules.
The entire system relies on small, single strands of DNA and three DNA polymerisation enzymes. When the 10 base DNA ‘prey’ strand finds a template strand, the DNA polymerisation enzyme increases the amount of prey in the ecosystem. However, these prey strands can then bind to ‘predator’ strands that are also complementary and slightly longer at 14 bases. Another of the enzymes then converts the prey–predator pair into two predator strands. To stop all the prey being turned into predators an exonuclease enzyme is also present, breaking up both predator and prey strands into single nucleobases that feed back into the system. The resulting rise and fall in populations mirrors ecosystem dynamics.
Rondelez is careful to stress that this is only a model of an ecosystem using the machinery of DNA replication, but ‘for some precise conditions, they are mathematically equivalent to predator-prey ecosystems’.
‘I think it is a magnificent achievement,’ says Erik Winfree of the California Institute of Technology, US, who develops genetic circuits and oscillators and describes himself as a friend and competitor of Rondelez. ‘There is theoretical evidence that information processing by simple biochemical systems can be surprisingly sophisticated,’ he adds, explaining that biochemical circuits could act as ‘brains’ for single-cell organisms. ‘The exploration of simple in vitro, cell-free biochemical circuits provides insights into this question.’
But while Fujii and Rondelez’s work demonstrates how a simple system can create a biochemical circuit, Rondelez doesn’t see DNA computation as a competitor to conventional computers. Instead, he suggests molecular computation is ideally suited for molecular applications such as smart drugs that include a diagnostic element to control their own release. But more work and better molecular resources will be needed, says Rondelez. ‘At the moment, I feel that we are just tinkering by reusing tools from molecular biology that have not been specifically designed for the purpose,’ he concludes. ‘I'm working on this idea.’