To insert these megachunks into the yeast chromosome, the team again uses homologous recombination, but with a twist. Each megachunk contains one of two selectable markers, URA3 or LEU2. The first megachunk might contain URA3, so successful integrants can grow in the absence of uracil. The adjacent megachunk contains LEU2. When it recombines into the chromosome, it deletes URA3 and inserts LEU2, such that the new clone can now grow in the absence of leucine, but not uracil.
“Basically,” says Boeke, “you can repeat this ad nauseum.”
Boeke and his colleagues have assembled an international consortium of researchers in the US, UK, and China — some with their own Build-a-Genome “franchises” — to use this process to tackle the yeast genome in its entirety.
But what makes Sc2.0 unique is that Boeke's group is not simply rewriting what evolution has written; they're tinkering, too. To reduce genomic instability, Sc2.0 will be stripped of both transposons and chromosomal tRNAs (the latter will be expressed from a plasmid instead). All TAG stop codons are being swapped with TAA, freeing up a codon for protein engineering with non-natural amino acids. Introns also are being removed, opening the door to determining whether splicing complexes do anything in the cell besides splicing.
Finally, every Sc2.0 gene is being tagged at its 3′ end with a loxP site to facilitate genomic “scrambling” in the presence of Cre recombinase, a form of controlled genomic mayhem that Boeke's design allows to happen precisely once in every cell's lifetime. He likens the process to “evolution on hyperspeed,” and when combined with selection procedures, the approach could result in yeast variants that may do remarkable things.
“It's a deck with 5000 cards. And now you can shuffle that deck any way you want,” he says.Shuffling the deck with MAGE
Researchers can also shuffle the genetic deck using multiplex automated genome engineering (MAGE).
Developed by Harvard University geneticist George Church, MAGE lets researchers make rapid, targeted modifications to a bacterial genome by introducing a pool of oligonucleotides targeting the desired loci into dividing cell populations every few hours; those oligos, which can specify specific mutations, insertions, or deletions, are incorporated into the lagging strand during DNA synthesis. It's like site-directed mutagenesis, but on a genomic level. “You can generate a cloud of different [genomic] solutions and find the right solution in there,” explains Church.
In a 2009 paper describing the technique, Church's team boosted lycopene production in Escherichia coli five-fold by targeting 24 biosynthetic genes simultaneously with degenerate oligonucleotides that either tweaked ribosome-binding sites or introduced nonsense mutations (5). Since then, Church's team has coupled MAGE with conjugative assembly genome engineering (CAGE) to essentially parallelize MAGE, using the combination to replace all 314 TAG stop codons in E. coli with TAA, and tied the technique to selectable markers (coselection-MAGE) to select for cells “that take up MAGE oligos and have a permissive replication fork in the desired region of the genome” (6, 7).
More recently, Church's team has enhanced coselection-MAGE with a pair of biochemical tweaks to the MAGE machinery — exonuclease knockouts (to preserve input oligonucleotides) and a “wimpier” primase — to boost efficiency by 16-fold (8). Turns out, they're going to need it. That's because the team is now applying this “souped-up” approach toward a monumental challenge: reassigning most of the genetic code's 64 codons throughout the bacterial genome, to make an organism that, by its very nature, is both resistant to viruses (because the two genetic codes would no longer be compatible) and yet amenable to protein engineering.
That's a genomic rewrite on a different scale altogether from Venter's Mycoplasma work or even Sc2.0, both of which were massive undertakings that nonetheless largely cribbed from the original genomic texts. What Church is doing is more akin to genomic encryption: the genetic words aren't changing, but to an outsider, they will appear to be gibberish.
“There are plenty of existing biotech companies that depend on E. coli for their livelihood, and E. coli gets infected by phages,” explains Church. “You can stick your finger in the dike for one or two of them, put a bandage on it, but to really have a general solution you need a new genetic code.”