The key words, though, are “should be”. For some researchers, the technology has been transformative. Model organisms that were refractory to basic genetic manipulations, including most vertebrates other than the mouse, were suddenly open to genetic manipulation. Sangamo BioSciences, a company that develops zinc-finger-protein-based therapeutics, has initiated human clinical trials with ZFNs targeting the cell surface receptor CCR5 to stymie HIV-1 infection (CCR5 is an HIV-1 co-receptor). And Sigma-Aldrich (which sublicensed the technology from Sangamo for research applications) has built libraries of ZFNs targeting every human, mouse, or rat gene ($4,000 each, according to the company's web site), or researchers can order custom ZFN for about $6,000 apiece.
But for other researchers, ZFNs have proved troublesome, being both expensive and tempermental. Voytas has a great deal of experience with ZFNs, having published some 20 papers on them and even co-founding the Zinc Finger Consortium to make the technology available to others. Yet, he says, “It's a little bit hard to predict how the zinc fingers are going to work in any given array. You can't just say, this finger recognizes GAT and it will always recognize GAT in every context I place it.”
In 2009, two groups independently discovered another useful class of DNA-binding proteins in the pathogenic plant bacterium Xanthomonas. The transcription activator-like effectors (TALEs) also bind DNA with a modular and predictable architecture. But instead of gripping nucleic acids in groups of 3, TALEs bind DNA through a long series of 34-amino-acid repeats, each of which contains a central amino acid pair that specifies a single nucleic acid base in the recognition sequence. By assembling TALEs with ordered arrays of these repeats, researchers could specify their binding sites with nucleotide precision. In 2010 Voytas and colleagues at Iowa State University coupled a custom TALE with the FokI nuclease domain to create a so-called TALE nuclease (TALEN).
Today Voytas mostly uses TALENs in his plant research lab. (He is also Chief Scientific Officer for Cellectis Plant Sciences, a company that develops TALENs for agbio applications.) In fact, they have become the tool of choice for many in the genome-editing community. Ekker, for instance, has been using them to toy with the zebrafish genome. “I have no zinc finger that competes with what are some of our worst TALENs,” he says. “And our best TALENs, there's no zinc finger or CRISPR system that competes.”
Ekker led one of two research teams that independently reported using TALENs to, for the first time, use an endonuclease to introduce novel DNA into the zebrafish genome by homology-directed repair. He and his colleagues used a novel high-efficiency TALEN system called GoldyTALEN, plus short single-strand oligonucleotides, to stitch in either a custom EcoRV or loxP site at specified locations within the zebrafish genome, in some cases with biallelic efficiency. “To our knowledge, these results represent the first description of successful [homology-directed repair, HDR] in zebrafish and the first demonstration of HDR using ssDNA as a donor template in vivo,” the authors write. (1) The second study, from Shuo Lin's group at the University of California, Los Angeles, used long double-stranded donor sequences to introduce GFP into three specific zebrafish genes in vivo, albeit at lower efficiency.
Their popularity notwithstanding, developing TALENs is not exactly a walk in the park. They are long and highly repetitive and their assembly can be challenging, says J. Keith Joung, an Associate Professor of Pathology at Harvard Medical School, who uses TALENs for zebrafish and human cell-based work. Joung's lab has developed three TALEN-design systems, including FLASH (Fast Ligation-based Automatable Solid-phase High-throughput system), a solid-phase method that uses libraries of preassembled cassettes of up to four TALE repeats to assemble the final molecule. Using this strategy, Joung's team rapidly built 48 TALEN pairs targeting EGFP and 96 pairs targeting endogenous human genes. Every one of the EGFP-targeted TALENs worked in human cells, as did 84 of 96 of the endogenous gene-targeting molecules. “We effectively don't think there are any real targeting range limitations to the TALEN platform,” he says.
An RNA alternative
What limitations, if any, will constrain the CRISPR/Cas system remain to be determined.