It used to be that if you wanted to make sophisticated genomic modifications in a model organism, you pretty much had to be working with mice.
First you would design a targeting vector, introduce it into mouse embryonic stem cells and inject those modified cells into a mouse blastocyst, which would take them up. Then came gestation, birth, screening, waiting for the desired pups to grow to sexual maturity, matings and crosses, more gestation, more screening and on and on.
Complex projects might take a year or more to complete. And it pretty much only worked in mice. For reasons not well understood, mouse embryonic stem cells seem to have particularly active homologous recombination systems. The same could not be said of, say, rats and humans.
Recently, though, new tools [1] have enabled researchers to make precision modifications in just about any organism they want, with nucleotide-level precision, incredibly rapidly. Most work by introducing a double-stranded DNA break at a specific location, which the cell then repairs. The differences lie in how the systems introduce the break, and the ease with which new sequences can be targeted.
ZFNs
The first genomic-editing strategy uses custom DNA endonucleases called zinc-finger nucleases (ZFNs).
Zinc fingers are transcription factors; each finger module recognizes three to four bases of sequence, and by mixing and matching those modules researchers can more or less target any sequence they wish (with some limitation: Sigma Aldrich, which commercializes ZFN technology, can produce a working ZFN for about every 50 bp on average, says Greg Davis, a Principal R&D Scientist who spearheads genome-editing technology at the company).
A ZFN is a heterodimer in which each subunit contains a zinc finger domain and a FokI endonuclease domain. The FokI domains must dimerize for activity, thus increasing target specificity by ensuring that two proximal DNA-binding events must occur to achieve a double-strand break.
The resulting cleavage event is what enables most genome-editing technologies to work. After a break is created, the cell seeks to repair it. The simplest method is “nonhomologous end joining” (NHEJ), in which the cell essentially polishes the two ends of broken DNA and seals them back together, often producing a frame shift. The alternative method is homology-directed repair. Here the cell tries to fix the break using another copy of the sequence as a backup—the other (unbroken) chromosome, for instance. By supplying their own template, users can force the system to inadvertently insert a desired sequence instead.
ZFN technology is owned by Sangamo BioSciences, which uses it to develop therapeutic products. For research applications, however, Sangamo licensed the technology to Sigma Aldrich.
According to functional genomics market segment manager, Shawn Shafer, the company offers both custom and off-the-shelf CompoZr® ZFNs for $4,000 to $7,000 apiece. These enzymes are all validated prior to sale, but recently, the company rolled out a “fast ZFN” option for those willing to skip the validation, providing four custom ZFNs for $3,000.
“We know that if we send four pairs of ZFNs for a knockout, 90% of the time at least one will be successful,” Shafer says.
Sigma Aldrich also recently launched a set of fluorescent protein-tagged ZFNs, so users can select (with flow cytometry) cells that are actually expressing the enzyme. The goal is to make downstream screening more efficient. “You treat the cells and then have to go through two to four weeks of single-cell cloning, and we are trying to help that process as well,” Shafer says.
TALENs
Though zinc fingers work, they’re relatively expensive and finicky to make, and researchers generally have to order them from Sigma Aldrich (though home-brew designs also exist, for instance at Addgene.org). Recently, however, a similar, but significantly more flexible system was discovered.
Transcription activator-like effector nucleases, or TALENs, are dimeric transcription factor/nucleases built from arrays of 33 to 35 amino acid modules, each of which targets a single nucleotide. By assembling those arrays just so, researchers can target nearly any sequence they like.
Stephen Ekker, director of the Mayo Addiction Research Center at the Mayo Clinic Cancer Center, compares TALENs and ZFNs today to transistors and vacuum tubes in the 1950s. “You could in principle design what we take for granted today [with vacuum tubes] but it would be very difficult; transistors made things much more accessible and straightforward.” Similarly, ZFNs provided the foundational proof of principle for genome-editing technology, but “TALENs have come in and do most of what ZFNs do, but cheaper, faster and better.”
Especially cheaper. Labs can build custom TALENs for a fraction of what ZFNs cost. Addgene sells individual TALEN plasmids for $65 apiece, and complete kits for a few hundred dollars. Dan Voytas’ popular Golden Gate TALEN 2.0 kit costs $425.
Voytas, professor and director of the Center for Genome Engineering at the University of Minnesota (and also consulting chief scientific officer at Cellectis Plant Biosciences, a company that develops TALENs for agricultural applications), develops strategies to optimize genome editing in plants. His university lab, he says, is “platform agnostic”—they use ZFNs, TALENs and the newly discovered CRISPR/Cas system (see below).
At present, Voytas says he prefers TALENs because his team has the most experience and success using them. But TALENs are also much larger molecules than ZFNs, he notes, and can be difficult to deliver efficiently.
To use his Golden Gate kit, Voytas says, users first identify the desired TALEN-binding site. They then select the plasmids in the kit for each required module, which can be cleaved and assembled in a step-wise manner in about a week.
The final and most difficult step is validating that the enzyme works as intended. “In certain systems, it’s the validation of function that can be the real sticking point,” he says.
If you prefer having your TALEN made for you, Cellectis Bioresearch sells both custom and premade enzymes. A nonvalidated, custom TALEN costs $3,360, and a design validated in human cells costs $5,000, says Jean-Charles Epinat, the company’s deputy CEO. Turnaround time for custom orders is about two weeks. Life Technologies also offers custom TALENs under its GeneArt® brand -- as well as TAL domains fused to activator or repressor domains for controlling gene expression.
CRISPR/Cas
The new kid on the block, genome-editing-wise, is the so-called CRISPR/Cas system. The subject of tremendous excitement in the research community—of 480 papers in PubMed on the topic, 160 have been published this year—CRISPR/Cas is like a DNA-targeted form of RNA interference.
In the CRISPR/Cas9 system (the unwieldy acronym stands for: “Clustered, regularly interspaced, short palindromic repeat (CRISPR)/CRISPR-associated-9”), the Cas9 nuclease makes a double-stranded break in DNA at a site determined by a short (~20 nucleotide) guide RNA. As with other systems, that break can be repaired by NHEJ or homology-directed recombination, depending on how it’s used.
Unlike ZFNs and TALENs, though, CRISPR/Cas is not human-made; the system provides bacteria a form of adaptive immunity. In August 2012, University of California, Berkeley, researcher Jennifer Doudna, with Emmanuelle Charpentier of the Hannover Medical School in Germany, figured out how the system works and showed it could be reprogrammed by swapping guide-RNA sequences. Since then, the field has exploded. (Science published an article on “The CRISPR Craze” in its August 23, 2013, issue [2].)
Search for CRISPR Vectors
Find, compare and review CRISPR Vectors
using the Biocompare vector search tool Search
That’s because CRISPR/Cas offers a number of advantages over ZFNs and TALENs: It’s simple, inexpensive, easily programmed and ridiculously efficient.
Doudna relates that one colleague told her of a researcher who was trying to do genome engineering in mice using TALENs. Trying seven enzymes targeting seven loci, the researcher after many months of labor was zero for seven, she says. “Within three to four weeks, the person in the lab doing the work was seven for seven with Cas9. And she’s never had a guide RNA that didn’t work in her mouse system.”
Harvard University geneticist George Church was one of the first to demonstrate the system’s utility in genome editing. “It came out of the blue for everybody,” he says. “It’s a real gift from biology.”
In particular, unlike TALENs or ZFNs, CRISPR/Cas can be multiplexed by adding multiple guide RNAs. Recently, for instance, Rudolf Jaenisch of Massachusetts Institute of Technology and the Whitehead Institute for Biomedical Research demonstrated that he could make five simultaneous mutations in mouse embryonic stem cells using five guide molecules [3].
Like TAL proteins, the CRISPR/Cas system can be tweaked with activator and repressor domains to control gene expression rather than editing. Yet questions remain concerning the system’s specificity (the guide-RNA sequence is shorter than most sequences targeted by ZFNs or TALENs, meaning the chance for off-target effects is higher). Several recent publications, including papers by Doudna and Church, demonstrate the potential for off-target effects is real, if manageable.
“I think what’s emerging is there will probably be ways to jigger the system,” Doudna says.
According to Doudna, for instance, off-target potential is influenced, among other things, by Cas9 concentration. And Church demonstrated recently that by using a mutant form of Cas9 that cuts only a single strand (as opposed to both strands)—a form called “nickase”—and two guide RNAs spaced closely together, it is possible to greatly increase the system’s cleavage-site fidelity [4].
“It’s like the difference between hybridization and PCR,” Church explains. “PCR is super-specific because it requires two independent hybridization events close together.”
Do-it-yourself CRISPR/Cas reagents are available at Addgene. Or, Sigma Aldrich just launched a commercial CRISPR/Cas implementation including a GFP-tagged Cas9 expression plasmid and a customizable guide-RNA sequence that users order via a web interface. Individual plasmids will cost $500 per target, says Shafer.
rAAV
If nuclease-based strategies don't grab you, genome-editing-wise, you can always try homologous recombination with recombinant adeno-associated viruses (rAAV). Unlike traditional homologous recombination, recombination induced by rAAV is relatively efficient in human cells.
“The mechanism is slightly different [than TALENs, ZFNs and CRISPR/Cas]: There’s no DNA damage induced. You’re simply asking the cell to take this piece of DNA and introduce that into the homologous region,” explains Eric Rhodes, chief technology officer at Horizon Discovery.
Horizon’s genome-editing platform (called GENESIS™) is built on rAAV, and the company offers a menu of some 500 ready-to-use X-MAN™ isogenic cell lines, each of which has in its genome a known mutation, generally mimicking that of a human oncology patient. Each line costs $990 for academic users. Alternatively, the company can generate custom genomic modifications for $35,000 and up. Academics can build custom lines themselves if they belong to Horizon’s Centers of Excellence program, through which the company has licensed rAAV technology for noncommercial applications. The company has established 30 Centers, including at the National Cancer Institute, Cambridge University, Johns Hopkins University and the Fox Chase Cancer Center.
Horizon Discovery also has indicated that it will be expanding its service offering to include other nuclease-based services; an announcement is expected soon.
In the meantime, if you’re itching to try some genome editing of your own, there’s no shortage of online help. A good first stop: genome editing expert Keith Joung at Massachusetts General Hospital lists a good set of resources on his website, including news groups for both TALEN and CRISPR/Cas technologies. After that, check out PubMed. At the rate this field is growing, who knows what you’ll find.
References
[1] Gaj, T, Gersbach, CA, Barbas CF III, “ZFN, TALEN, and CRISPR/Cas-based methods for genome engineering,” Trends Biotechnol, 31:397–405, July 2013. [PubMed]
[2] Pennisi, E, “The CRISPR Craze,” Science, 341:833-6, August 23, 2013. [PubMed]
[3] Wang, H, et al., “One-step generation of mice carrying mutations in multiple genes by CRISPR/Cas-mediated genome engineering,” Cell, 153:910-8, May 9, 2013. [PubMed]
[4] Mali, P, et al., “CAS9 transcriptional activators for target specificity screening and paired nickases for cooperative genome engineering,” Nat Biotechnol, August 1, 2013, doi:10.1038/nbt.2675 [PubMed]
Image: CRISPR/Cas9 system, courtesy of Sigma Aldrich.