By any account, the start of 2013 was extraordinary for would-be genome engineers. Between January 3 and January 29 six papers appeared in the pages of Science, Nature Biotechnology, and eLife, all leveraging a curious form of bacterial immunity called CRISPR/Cas to nip and tuck the genomes of human and mouse cells, bacteria, and zebrafish.
CRISPR/Cas is as essentially a bacterial adaptive immune system enabling cells to “remember” foreign nucleic acid invaders and silence them by site-specific DNA cleavage. Jennifer Doudna, a Howard Hughes Medical Institute Investigator and Professor of Biochemistry and Molecular Biology at the University of California, Berkeley, with Emmanuelle Charpentier at the University of Umea in Sweden, unraveled the biochemistry of the process in late June 2012 and showed the system could also be reprogrammed to recognize different targets in vitro.
Clearly, the genome editing community was watching: Five other groups (plus Doudna's) quickly picked up the ball and ran with it, demonstrating that CRISPR/Cas systems could create targeted double-stranded breaks in genomic DNA in live cells, the first step in genome editing. Manuscripts rolled into Science's editorial offices a scant three months later.
“There's a real hunger in the scientific community for a tool like this,” Doudna concludes.Building on cellular tools
To appreciate the advance CRISPR/Cas represents, it helps to have some historical perspective. In the biological pleistocene—that is, prior to the mid-1990s—few options existed to manipulate the genome of a cell or organism, and those that did involved assembling a sequence in vitro, introducing it into cells, and hoping for homologous recombination to occur. There was no way to direct the process really, because there was no way to force the cell to take up the sequence of interest and copy it into its genome.
Today, that limitation is largely ancient history thanks first to custom DNA endonucleases (zinc-finger nucleases, TALENs, and meganucleases) and, more recently, CRISPR/Cas. As Dan Voytas, Professor and Director of the Center for Genome Engineering at the University of Minnesota, explains, these systems all work their magic by tapping into the same fundamental molecular process.
“The bottom line is that they are all ways to create a targeted break in a chromosome in living cells,” Voytas says. “That seems like a strange way to begin your precise modification, but once the chromosome is broken, [the cell] seeks to repair the break, and then we can direct how that break gets repaired.”
The cell uses—and genome engineers exploit—two primary mechanisms to repair double-strand breaks. The simplest and most efficient strategy is non-homologous end-joining (NHEJ), in which two DNA ends are simply glued back together. If that process worked perfectly, there would be little value in NHEJ as a genome-editing tool, but as it turns out, the process is error-prone. Bases are often added or deleted to produce small “indels” and, likely, either frame-shift or nonsense mutations within the coding sequence of a gene. The result: a quick-and-dirty gene knockout, no embryonic stem cells required.
“I call that the ‘laser method’,” says Stephen Ekker, Professor of Biochemistry and Molecular Biology at the Mayo Clinic Cancer Center, who uses custom endonucleases to manipulate zebrafish genomes. “We're using a guided laser where you zap the DNA and then, poof! Some changes happen right there.”
The second repair strategy is homology-directed repair, the equivalent of the cell going to its backup hard drive to restore a damaged file. Normally, that correct information is contained on the second chromosomal copy. But genome engineers can hijack that process too if they supply “donor” copies of their own containing anything from a point mutation or two to an entirely new gene sequence.Designing custom endonucleases
So, how does one generate the initiating DNA lesion that induces these mechanisms? Until earlier this year, the only option was to build a custom restriction enzyme.
Researchers at Johns Hopkins University first developed this idea in the mid-1990s by coupling the endonuclease domain of the FokI restriction enzyme to the DNA-binding domain of zinc-finger transcription factors, creating what have come to be called zinc finger nucleases (ZFNs).
It turns out that zinc-finger transcription factors recognize DNA in a very modular fashion with each “finger” binding a specific and predictable trinucleotide sequence. This means it should be possible to target any desired sequence by simply assembling the proper set of fingers, three bases at a time, and expressing the resulting protein in cells.