If there's one reason to be worried about genetic engineering, it's that life doesn't operate in a vacuum. Just as a substitute beekeeper released what would turn out to be a continent's worth of insect genetic material, genetically modified organisms (GMOs) are not guaranteed an inert or even well-contained environment.
GMOs are now deployed in the wild, as crops; in fermenters and even drug-producing (Atryn) goats; and in ourselves, as live vaccines (and soon probiotics). We're on the cusp of becoming GMOs as well—a clinical trial recently demonstrated safety and efficacy of a gene therapy for RPE65-associated retinal degeneration.
Free Trade In Situ
Organisms that result from gene modifications are very similar to their unmodified parents, to the point of being able to easily exchange genetic information with them. One potential problem with easy exchange of genes is that it wreaks havoc with your GMO. So, while you might have engineered your bacterium to be dependent on the addition of some particular nutrient (auxotrophic), it might also have a large pool of neighbors who can offer genes to correct that deficiency.
It's possible to reduce this risk by knocking out recombinogenic proteins in your organism, but even then, other risks remain. In
2009, viral contamination brought down the bioreactors at Genzyme's Allston plant, costing millions in lost revenue. The opposite danger also exists: other organisms gaining the engineered genes.
The power of genetic engineering is in its ability to precisely transfer genes into unnatural contexts—for instance, taking a bacterial gene and putting it into a plant. But making these genes work in their new hosts usually entails adding promoters and
altering codons, effectively putting the gene into a new realm of life. So what was a bacterial gene is now, engineered, a plant gene, and as such it can be shared among compatible plants. Licensing becomes a little more challenging when the gene spreads on its own. Also, this allows for combinations of genetic material that Nature hasn't had the opportunity to try, with unpredictable and potentially dangerous effects.
One option for preventing undesirable interactions is to make the organism (particularly its genetic information) chemically unrecognizable to other life and vice versa. This mutual incompatibility (and ideally inactivity) is termed bio-orthogonality, and it could function as an excellent partition between GMOs and their environments.
Looking for orthogonality is like asking, “What kinds of variations are permitted in the chemical conventions of life? Can we create organisms based on unnatural nucleic acids, amino acids, enzymes, metabolites, and so on?”
Genetic bio-orthogonality might just be feasible. Last year, Craig Venter announced the first "synthetic cell," actually a demonstration of inserting a synthetic genome into an already-functioning cell. This summer, another group demonstrated a bacterium evolved in the lab to preferentially use chlorouracil in its DNA in place of thymine. Taken together, these achievements look like milestones on the way to an organism that stores its genes in such a way (unnatural nucleic acids) that no other species can make use of its information and vice versa.
But unnatural nucleic acids aren't quite that easy: Chlorouracil can also be mistakenly integrated into mammalian genomes, with mutagenic side effects. This raises several points: The orthogonal organism should not produce toxic wastes or require toxic (or expensive) compounds to grow, and it should not produce toxic degradation products if killed. However, compounds that are very similar to existing metabolites are likely to be naturally functional or poisonous inhibitors/mutagens.
Modifications to the DNA backbone—using a hexose, “locked” ribose, or threose instead of ribose, for example—have been shown to be possible, but differences in flexibility, base-pairing, and melting temperature make for a lot of potential snags at the genome level. Furthermore, the more orthogonal you get, the more enzymes you have to engineer; not only do polymerases need to be altered but also ribosomes, tRNA synthases, nucleases, and maybe even kinases and other enzymes. The chemically incompatible life form is quite a piece of work.
A more reasonable approach is to meet halfway on the orthogonality. Instead of making the nucleic acids themselves orthogonal, make the code they carry orthogonal. Switching GGN from glycine to, say, proline, and so on is entirely possible with the “synthetic cell” mentioned earlier; just change all of the coding sequences to match the altered tRNA and jump right into troubleshooting! Such a system gains most of the benefits—like viral resistance and reduced gene flow—with far less enzyme engineering.
Even so, the demand for such a platform is far outweighed by the costs of making it—for now. But as gene synthesis becomes cheaper and more of the development process becomes automated, the orthogonal organism could become a reality. As difficult as building the fence might be, making the neighbors
well-behaved is even harder.