Farren Isaacs’ Recoding of the Genetic Code and the Future of GMOs

Yale Professor Farren Isaacs’ big moment on late night TV came on October 24, 2013. After a night out with the lab celebrating their creation of the first genomically recoded organism (GRO) in the lab, Isaacs was inundated with messages from friends and family — his work had been spoofed in a sketch on Jimmy Kimmel Live. As the lab-coat-clad professor “Faren Isaacs” (missing an “r”), touts the significance of his work and reassures Kimmel that the newly created life form was harmless and entirely safe, the sounds of screaming grad students and breaking glass filtered in from off-camera. The situation deteriorates, and “Faren” is soon quivering and holding a sign saying “HELP ME” in dripping red letters as ominous shadows and tentacles destroy the camera. The fictional Professor Isaacs is, presumably, consumed as we cut back to Jimmy Kimmel’s studio.

The real Professor Isaacs says he “got a kick” out of the sketch. But in fact, the same breakthrough which Isaacs was celebrating that October may hold the key for ensuring that our fears about genetically modified organisms (GMOs) do not come to pass. While genetically modified bacteria sitting in reactors to make medicine and other chemicals aren’t going to turn into tentacled beasts and devour their creators, as we look at possible medical or environmental applications for such GMOs in open systems, it is prudent to try to prevent them from escaping into the wider world. The GMOs Isaacs works with have a special sort of leash: instead of inserting or removing a few genes, like editing the words in a book, Isaacs has changed how the cells translate the genetic sequence, changing the language the book is written in.


When a GMO is recoded, all the TAG stop codons are changed to TAA stop codons via MAGE so they still stop normally. Foreign DNA, with the TAG codon, will not translate correctly.

The genetic code underlies the “central dogma” of molecular biology: our cells “read” three-letter sequences of DNA, pair each sequence with one of the 20 natural amino acids, and join these amino acids together to form a protein. (The resulting proteins, in turn, do basically everything else.) In most organisms, when a cell comes across sequences TAG, TAA, or TGA, it stops building the protein and releases it to do its protein duties. But in Isaacs’ recoded E. coli, which he commenced as a postdoc with Professor George Church at Harvard, the TAG codon doesn’t end translation – instead it adds a synthetic amino acid (sAA), and continues translating. This expands the genetic “vocabulary” of the organisms (which can now use more than just the 20 amino acids) and changes when proteins stop transcribing. Such genomically recoded organisms (GROs) could play a starring role in the development of safer, and more widely deployed genetic technology over the next several decades.

For traditional biological containment, scientists engineer organisms to depend on essential chemicals by knocking out the genes that produce those molecules – the cells must then be fed these molecules, or die. But, in the words of Dr. Ian Malcolm in Jurassic Park, “life, uh, finds a way.” In this case, that way is horizontal gene transfer. Microorganisms swap genes with each other, so if we’re using them outside isolated reactions, bacteria can readily pick up missing genes, counter their engineered deficiencies, and escape into the wider world. Even with engineered dependence, about one in every million or so cells overcome their “addiction” – and, while this doesn’t sound like much, incubators can grow hundreds of billions of cells so escape becomes almost certain. By changing the language of DNA, though, Isaacs hopes to block such “communication” between organisms, cutting off this escape route: if the GRO picks up a gene which would otherwise allow escape, but which ends in a TAG codon, the ribosome will just add an sAA and continue translating, resulting in an aberrant protein and preventing escape.


The MAGE device Farren Isaacs' lab uses to assemble genomes with numerous changes, such as replacing all TAG codons with TAA. Using MAGE, Isaacs and his lab can do automatically what formerly took hundreds of man-hours.

To get this to work though, Isaacs had to change all the TAG stop codons in his GROs to TAA – otherwise the cells wouldn’t be able to correctly complete any protein whose gene ends with TAG, and wouldn’t be useful in any environment. Beginning while he was working at Harvard under Professor Church, and continuing at Yale, Isaacs created genome engineering technologies (MAGE: multiplex automated genome engineering, and CAGE: conjugative assembly genome engineering) to drive large-scale genome modification. MAGE enables editing bacterial genomes at a wide swath of points, selecting bacteria with the correct modifications, and combining correct modifications using CAGE until the entire genome had been scrubbed of TAG codons. Isaacs describes MAGE as a word processor for genomes, allowing us to delete and change genes on the genome with unparalleled accuracy and precision – and because it’s entirely automated, it speeds up the editing process as well.

With the cells’ language changed, Isaacs’ lab found ways to exploit the expanded vocabulary. The lab put TAG codons into other positions on the genome, incorporating sAA’s into essential proteins. This provides another barrier to escape – if the bacteria aren’t fed the artificial amino acid, they can’t produce essential proteins, and die. By incorporating the sAA’s into important parts of multiple proteins, Isaacs brought the escape frequency down to undetectable levels (or less than one cell in a trillion) in their laboratory experiments.

Is that enough? On one level, it depends on the application. One potential use of live GROs is in vaccines, full of bacteria which imitate the pathogen we want to immunize against – such injections would require “only” millions of cells, so would be easily contained by these methods. Another, though, is environmental remediation, where we could employ more than trillions of toxin-eating cells over a large site. Then, even at the levels Isaacs has achieved, escape may be a possibility – a possibility that can only be quantified through larger-scale testing of the lab’s containment strategies.

Fortunately, specialized GROs tend to be fairly fragile, and not likely to survive robustly in competition with natural species. So even in the case of escape, Isaacs tells me, “I’m not really concerned,” but adds “I just think it’s important to proceed safely.” So he continues to work towards better methods – a recent paper in Nature Biotechnology details his lab’s success in incorporating even more sAA’s into bacterial proteins. Poor Faren Isaacs from Jimmy Kimmel may have much to fear from GRO’s, but thanks to Farren Isaacs from Yale, the future of bioengineering looks brighter and safer.

About The Author

My time in Princeton is split between Frick labs and running around breaking things (and eardrums) with the Band. Outside the bubble, I hike (OA and otherwise) and write the occasional ridiculous poem.