Analysis: The gene machines
Bio-engineering is in the news, as readily-available kit helps researchers attempt to tame life, but complexity still dogs the would-be synthetic biologists.
In promoting his laboratory's work, J Craig Venter has claimed that the technology of synthetic biology can arrest climate change. Too much carbon dioxide in the atmosphere? Let specially designed bugs eat it. Not enough fuel? Bacteria can synthesise it.
But Venter is far from alone. The technology needed to redesign living organisms is so readily available that it is becoming a teaching tool for undergraduates. The equipment and materials are cheap enough to put into the hands of students who, instead of having to learn about the mechanics of life from highly regimented experiments, can have a hand in building their own brand of organism.
The focus for the undergraduate work in synthetic biology is a competition that has been hosted by the Massachusetts Institute of Technology (MIT) since 2005. The iGEM competition challenges student teams to develop genetic machines using standard components and operate them inside living bacteria.
Some projects have looked at whether bacteria can function as rudimentary computers, fitted with genetic logic gates that switch in response to signals. One MIT team created a sweeter-smelling form of Escherichia coli - a bug with a foul odour that's more familiar as a source of food poisoning. Not only that, but it could switch between minty freshness and the perfume of bananas.
The genes used to program the bacteria's behaviour form a simple logic circuit, using chemical signals to indicate what they can sense. While the bacteria are in their growth phase, they exude methyl salicylate - the essence of wintergreen. But as they fill their culture plate and run out of nutrients, they switch to making isoamyl acetate - the smell of ripe bananas.
Professor Alfonso Jaramillo of France's Centre National de la Recherche Scientifique (CNRS), has gone further by designing a JK-latch - a memory circuit familiar to electronics engineers - but based on bacterial cells and using genetic switches in place of transistors.
"Bacteria certainly can be seen as genetic machines. You have sensors, internal logic and modularity," James Brown, a graduate student in Professor Jim Haseloff's team at the University of Cambridge in the UK, said at a recent seminar.
Are they truly genetic machines? "It would be good to have components that could be provided and reassembled: literally plug and play the way you can with an electronic circuit or a mechanical circuit system," said Brown.
Some elements of the iGEM competition make synthetic biology look like engineering. Teams are encouraged to use genes stored in the database operated by the BioBrick Foundation - a Cambridge, Massachusetts-based group formed by researchers from MIT, Harvard University and the University of California at San Francisco.
The BioBrick Foundation describes its repository as "open source, standardised snippets of DNA that encode basic biological functions". Brown said each segment of DNA has a datasheet, "just like any electronic component".
If there is no BioBrick, researchers can turn to one of the custom-gene services that have sprung up in recent years. Jeremy Minshull, president of DNA 2.0, based in Menlo Park, California, said: "Most of what we do is make something that looks very much like a natural gene. People write to us and say they can't get hold of one and we make one for them."
Unfortunately for would-be bio-engineers, assembling your custom bacteria is not as simple as picking genes from a list, stitching them together and inserting them into the host bacterium.
"We are not there yet," said Brown. "There is no rigorous way for constructing devices yet. It is not yet reminiscent of mechanical engineering."
The problems begin with the production of proteins from the recipe specified in the genetic code. What biologists know is that the base unit of information in the gene is a group of three nucleic acids in a sequence. At each point, there can be one of four different kinds of nucleic acid. That gives a total of 64 combinations. A ribosome reads along the length of the gene, using the codons to pull different amino acids into the chain it is building to produce a protein. That protein will then perform some chemical function, such as combining smaller molecules to produce wintergreen. "The protein is the part of the biological system that does most of the work," said Minshull.
In principle it is simple. But there are issues even here.
Minshull said: "There isn't one to one correspondence between codons and amino acid." There is a lot of redundancy in the codon codes, and there are interactions that mean you cannot reverse-engineer a gene from the resulting protein. Researchers are trying to work out what the interactions are, using computer-based analysis to see which sequences are the most efficient at generating proteins in vivo. "In experiments, you can get two orders of magnitude difference just by changing the codons," said Minshull.
Biologists are trying to work out whether they can simplify matters by introducing restrictions. Venter is working with as simple a genome as possible. Others are looking more closely at how proteins and other molecules work in vivo.
Professor Wendell Lim's group at the University of California is using synthetic biology not to synthesise materials but to investigate the role of evolution in the development of cells. "We view synthetic biology as a complement to discovery biology," said Caleb Bashor, a graduate student in Lim's group.
"We are looking at how cells process information. What is the logic? What are the rules? And how can we use the principles we think we understand to build novel signalling networks?"
The biggest problem is crosstalk. The presence of a chemical that should trigger a reaction involving just one protein turns out to initiate others. However, it is possible to limit the interactions by building all of the pieces needed to manufacture a chemical or react to a signal on one large protein: a scaffold.
"It turns out that signalling proteins are highly structurally modular," said Bashor. They sit on a scaffold. Simply replacing and rearranging elements on one scaffold can 'reprogram' the protein. "It suggests that the complexity of networks is due to the interactions. But the question still remains: are the elements functionally modular?"
Based on the work performed in the last ten years, Bashor said: "Scaffolds appear to provide a powerful platform for exploring the plasticity of [cell] pathway signal processing: a tool for understanding circuit-design rules. Using them, we could rewire cells with new or modified behaviours."
Such is the complexity of the systems, some synthetic biologists question whether Venter is on the right track in attempting to create a novel bacterium with the smallest number of genes to be viable when nature seems to be able to provide workable templates for their experiments. Restricting the number of genes may not make all that much difference to dealing with the complexity of interactions, crosstalk, protein folding and the effects of small changes in the environment on gene expression.
Even when a modified organism seems to produce the correct results, it might not work on location. "A glass screen is not maize or a mouse," said Minshull.
It will be many years before biologists can use computer models to predict accurately how a modified organism will behave. However, their ability to produce viable results with comparatively few resources has convinced biologists that the synthetic form is here to stay.