A new class of species may have been invented, and all in the name of safety. Colonies of Escherichia coli—the gut bug famous from food poisoning outbreaks—have had their genetics tweaked in a way that prods them to produce useful molecules, such as fuels or pharmaceuticals.
 
But such modified microbes might prove problematic if the microscopic bacteria escape the lab, especially if researchers' goal of endowing them with resistance to viral infection via genetic manipulation is achieved. Such resistance is useful for keeping colonies alive for research but could turn a novel life-form from a microbe that can only survive in the lab or an industrial setting into a bacterium that could outcompete its wild brethren. So now researchers at both Yale and Harvard universities have demonstrated how to build in safety controls: They have made the microbes dependent on artificial amino acids to make the proteins necessary for life. Unless they are purposely fed with those amino acids, any escaped bacteria would die. "This adds another important safety barrier," says biologist Farren Isaacs of Yale University, senior author on one of two papers laying out this approach published online in Nature on January 21. (Scientific American is part of Nature Publishing Group.)
 
Teams at Yale and at Harvard Medical School worked with so-called genetically recoded organisms, an E. coli that has had genetic modification taken to the next level with multiple changes to its genetic code. In fact, these recoded E. coli (RE. coli) have had one out of its genome's 64 codons—tiny phrases of DNA or RNA that correspond to a specific amino acid or identify when cellular machinery should stop building a protein—swapped out wherever that snippet of genetic code appears throughout the entire 4.6 million base pairs of DNA in its genome, more than 300 modifications in all. As a result, this RE. coli cannot build on its own at least one of the proteins necessary for its survival without a synthetic amino acid of unusual size and shape not found in nature. In essence, the researchers have built the first life-forms to rely on a 21st amino acid, outside of the 20 natural amino acids employed to build life-sustaining proteins by all known life on this planet.
 
There are many such synthetic amino acids to choose from: The Harvard researchers built in reliance on a synthetic amino acid called bipA (short for biphenylalanine); the Yale team chose the synthetic amino acid pAzF (4-p-Azido-L-phenylalanine). Several genes in each version of RE. coli that control essential protein-making were tweaked so that assembling the protein required the synthetic amino acid rather than one of the 20 common amino acids. "If they can't produce it [the synthetic amino acid] and they can't scavenge it, then they die," explains geneticist George Church of Harvard Medical School, senior author on the other paper, "except for the ones mutated enough so they no longer need it."
 
In the future, the genes for all 290 proteins that E. coli produces could be tweaked so that an artificial amino acid is needed to generate the encoded protein, although Church suggests that just making three or four of these proteins reliant on a synthetic amino acid should be sufficient to ensure a less than a one-in-100-billion chance of a random mutation enabling some version of the tweaked microbe to live outside the lab. Those odds would not mean escape is impossible, of course, but "impossible in the real world means that practically the probability is lower than any size population that you would grow," Church notes. Other escape routes could be RE. coli cells that scavenged leaked amino acids from other dead cells or swapped enough genetics with other microbes to find a way to survive. Current National Institutes of Health regulations for biotechnology require that no more than one such mutant appear per 100 million cells.
 
The new, built-in safety control—such amino acids are not found outside the lab—also could come with a bonus for industry: the bacteria’s resistance to infection by viruses. The earliest versions of recoded E. coli had fewer codon changes but these tweaks conferred immunity to a common virus. It is that ability to resist viruses that may cause industry to take up this new genetic safety code as well. "Once you combine those two, then it becomes very attractive," Church says, because the microbes are both safely contained and have less risk of being wiped out by an infection.
 
"People are so worried” about the safety of genetic engineering, says Cai "Patrick" Yizhi, director of the Genome Foundry at the University of Edinburgh, who is working on similar genetic containment strategies for yeast. “You can rewrite the code of life," but the new bacteria, he suggests, will be safer for reasons beyond their dependence on synthetic amino acids. The new species, he says, will have trouble surviving in the wild because they will not be able to swap useful genes with other bacteria readily. “They cannot communicate with the wild type,” he notes, “because [the researchers] have engineered a species that speaks a different chemical vocabulary."
 
Genetic control
In 2008 the Cambridge, Mass.–based company Genzyme had to shut down its bioreactors in Belgium, where Chinese hamster ovary cells churned out the drugs Cerezyme and Fabrazyme, which treat genetic diseases. The ovary cells had fallen prey to an infection of vesivirus, which impeded their growth. The following summer Genzyme experienced the same viral infection at its bioreactors in Allston, Mass. Each facility had to be shut down for a few months to allow for complete cleaning, which resulted in more than $100 million in lost sales of the drugs.
 
Such bioreactors are big business today, used throughout the world to produce everything from insulin to dairy products such as yogurt and cheese. Although the fermentation occurs in physically sealed containers, those seals are not perfect and contamination can occur where viruses get in and ruin the process, as in Genzyme's case. Or industrial microbes could escape and, if genetically superior, replace the wild versions.
 
Typically, industrial microbes, such as the modified E. coli DuPont uses to produce propanediol, a precursor to plastics, cannot compete against their wild cousins if released. But if the microbes had genetics that made them resistant to viral infection, that resistance could prove a compelling edge in the evolutionary battle for survival. "That category could outcompete and become an invasive species in the wild," Church suggests. "It doesn't mean it would be bad for the environment, but it could be impactful."
 
To get ahead of that potential problem, the teams tested to see whether the need for synthetic amino acids would prevent the virus-resistant bacteria from living outside the lab. "You want it genetically and metabolically isolated from other species," Church notes. His group grew a trillion of the RE. coli cells dependent on synthetic amino acids to see if any could evolve to escape, and none did. In the experiments where RE. coli did manage to survive through mutation, regular E. coli easily outgrew the escaped mutants.
 
RE. coli's genetics are so different from its wild cousins, in fact, that it should not be able to successfully swap genes with them. The RE.coli changes are dispersed throughout the genome; therefore, either not enough genes come over in such a swap—known as horizontal gene transfer—to allow proper protein manufacture and survival or the bugs trade so much genetics that the tweaked E. coli loses its other tweaks, too. In the work by the Harvard group, "either they didn't take in enough DNA to escape or they took in so much DNA that they overwrote the whole genome," explains Daniel Mandell, a geneticist at Harvard Medical School and lead author on the paper from Church's lab.
 
Essentially, RE.coli speaks a different chemical and genetic language than E. coli. "Recoding establishes a genetic firewall that prevents the exchange of genetic material," Isaacs says. His lab is also working on other methods of such containment, including further tweaking of how individual genes are expressed or creating a piece of cellular machinery that cuts the genome if certain chemicals are no longer present.
 
The most significant barrier to adopting RE.coli and other recoded organisms is likely to be the cost of the safety measures. For example, pAzF costs nearly $300 per liter of culture to add. Given that an industrial fermenter holds at least 1,000 liters, that cost could prove prohibitive. On the other hand, bipA costs roughly $4 per liter. "These amino acids were chosen to be as inexpensive as possible," Church says. "Really chosen to be usable at low concentrations and low cost."
 
But such high levels of safety control raise the prospect that such recoded organisms could even be released into the wild one day. "Introducing these types of safety measures sets the stage for new applications of these organisms beyond some of the contained fermentation processes," Isaacs says. "You could imagine using specialized microorganisms for bioremediation or engineered organisms as probiotics that can combat disease."
 
Church's lab is now extending this work to change seven codons across the entire RE. coli genome, rather than just the one, which could make for a microbe that can produce a lot of a given compound, resist viruses and be incapable of escape from the lab or causing trouble if released in the wider world. "I have been pleasantly surprised at how low the impact has been on productivity," Church says.
 
And Isaacs notes that incorporating new, synthesized amino acids into bacterial proteins may make new chemistry possible, by altering the function of these proteins. That functional change, in turn, could allow the tweaked RE. coli to become living foundries that produce new types of products or materials combining both synthetic and natural amino acids, such as new types of protein-based drugs or new plastics. "It could be the foundation of new types of drug delivery vehicles or nanostructures as well as antimicrobial pesticides," he says. "Early results are encouraging."
 
New life
The significant genetic tweaks also mean that this new synthetic RE. coli is not just a new species. It may merit designation as a new class or even a new kingdom of life because it can no longer interact genetically with other microbes. "Once this gets different enough, it’s a genetic code barrier unlike any that has ever existed in the kingdom of life," Church argues.
 
That said, escape from this genetic shackle is not impossible. Modifications that improve life, whether conferred via random chance or human tinkering, will prosper and no safeguard has yet been invented that has been able to stop evolution finding a way around it. "It's a great technology but it's not the end of the game," Cai notes. "We are just in the beginning of a continuous battle."
 
And the work does not stop with the E. coli. Such recoding could be scaled up, with challenges of course, to bigger industrial organisms, such as yeast or even, maybe one day, plants. "Plants have nine times as many protein-coding genes as E. coli," Church says, noting that this difference means more DNA has to be synthesized and inserted into the plant genome while also coping with the relatively slow growth of plants or plant cells compared with microbes. "It's more of a challenge but it's not out of reach."