If there is a difference between synthetic biology and plain old genetic engineering, it is one of scale. Genetic engineering typically relies on just adding a gene here or there. Synthetic biology is about working on big chunks of the genome - inserting or deleting entire sections.
As a result, it's not a surprise that people see gene synthesis as a key enabling technology for synthetic biology. You need bigger chunks of DNA. Ergo, you need to synthesise a lot of DNA. The J Craig Venter Institute (JCVI) is the most famous proponent of this philosophy having enlisted the help of four gene-synthesis company, and a lot of delicate cloning work at the JCVI, to build an entire genome from scratch. Ultimately, that synthetic genome is meant to reboot an existing bacterial cell with a new operating system. But it's by no means a popular approach, not least because it's so expensive to do and because it's not that useful in practice.
At Synthetic Biology 4.0, some fretted about the massive cost of synthesising a genome-sized lump of DNA and how it would prove a roadblock to serious synthetic biology work until an equivalent of Moore's Law brought the cost down. However, few of the practising biologists at SB 4.0 were all that bothered. In one panel, one wag quipped was that there was just one customer for whole-genome synthesis: the JCVI.
The issue is that you would only use whole-genome synthesis in one instance: when you absolutely know that's the sequence you need. The trouble is, right now, that's very rare. The JCVI knows the sequence it needs for its synthetic genome. Because, with a few modifications to encode a message in the DNA, it was copied wholesale from the natural genome for Mycoplasma genitalium.
Until computer simulations of the processes inside a cell get so good that you can predict everything, evolution is going to be a vital process in engineering biological organisms (and I'm aware I'm making the assumption that that will happen...eventually). You simply don't know the DNA sequence until you've tried out a few options and worked out which ones don't work or do but simply get evolved out.
In the case of the work just published online at Nature by Harris Wang and Farren Isaacs, working in Professor George Church's lab at Harvard Medical School, why not try 4 billion options? And the next day, another 4 billion.
Isaacs calls the approach a "synergy between engineering and evolution". The engineering part is knowing which genes or segments of DNA you are going to hit. The evolution part comes down to creating many, many mutants that vary only slightly but, because gene networks have so many links, can give rise to dramatic changes in performance.
The aim was to increase the production of an chemical, lycopene, by E coli. Used to make the red pigment carotene and similar chemicals, lycopene is an isoprenoid - a family of chemicals that Jay Keasling's Amyris Biotechnologies over in California has been working with as a source of biofuel. Within five days of producing many different variants, they isolated one that produced five times more than the naturally occurring species that they started with.
One reason why whole-genome synthesis for this kind of work is a poor fit is that the genome changes only slightly. The researchers identified 24 sites worth hitting across the entire E coli genome, which contains some 4000 genes. The changes involved replacing a section of less than 100 nucleic acids at each target - and actually changing only around 30. So that's fewer than 1000 base pairs out of a total of 4.5 million - a mere 0.02 per cent of the entire genome, but they are spread all the way around the chromosome. In that situation, it makes far more sense not to rewrite the entire genome but perform a series of edits.
The neat thing about the the Multiplex Automated Genome Evolution (MAGE) technique is that you don't have to do each edit at a time. By taking advantage of the way that a virus inserts DNA into a genome, it's possible to simply mix all of the replacement sections of DNA (oligos) with the cell cultures and then use electric shocks to open up the cell wall temporarily to let the oligos in. Not all of the oligos take in all the cells - you get a lot of variation, especially if you do what the team did and deliberately design a number of very similar oligos to target a particular site. However, if you run enough cycles, the method will see the cells converge on a new genome that contains a particular set of replacements. This is a handy aspect of MAGE that is being used by Isaacs on a larger programme to rework the core genetic code itself.
However, if you run a smaller number of cycles, you wind up with a lot of mutants that you can then culture to see how the perform at chemical production. It's effectively a fast-forward button for directed evolution and one which can explore many more options than simple natural variation, which normally only goes through single genetic alterations, would find.
In common with a growing number of labs, automation is a key part of MAGE, although it is by nature a highly parallel operation: Isaacs says it should be possible to develop techniques that can target key sections of all 4000 E coli genes at once. A lot of those mutants will not survive but, as work on a smaller by Mark Isalan and Luis Serrano and others at Barcelona found, genetic rewiring of this kind can yield surprising and potentially useful results.
However, the technique is already semi-automated and the Harvard team is working with some industrial partners to develop a more robust system with an eye on commercialisation.
At the Weizmann Institute in Israel, Professor Ehud Shapiro's team has been working another approach to implementing large-scale changes to genomes without actually constructing them from scratch. Shapiro's "DNA word processor" uses the classic PCR reaction to copy and edit sections of DNA but, thanks to robotics, on a massive scale. (The video is accelerated but probably still a dream for put-upon postgrads who have to do this kind of stuff by hand). Although the technique can produce a single edited long strand of DNA, it is also good at generating subtle variants. Several researchers have asked Shapiro to build libraries of genetic sequences for them so that they can try the variants out in parallel to see which ones work and which don't. Try doing that with classic gene synthesis.
Isaacs says he sees large-scale gene synthesis and techniques such as MAGE being complementary. Talking to Brandon Keim of Wired, Church seemed less confident about the future of genome synthesis: "There are very few clearly articulated examples where anyone needs to change more than a few dozen genes or base pairs."