NEW YORK (GenomeWeb) – Scientists from the University of Massachusetts Medical School have engineered a new way to do CRISPR/Cas9 genome editing by fusing a mutant Cas9 with earlier technologies used in the field, such as zinc finger proteins.
Led by senior author Scot Wolfe and first authors Mehmet Bolukbasi and Ankit Gupta, the researchers published their proof-of-concept study last week in Nature Methods.
"The things we like about this system is that in principle it's very flexible," Wolfe told GenomeWeb.
The scientists changed the DNA binding efficiency of Cas9 by mutating it, but physically linked it to an extra DNA binding domain to make up for that deficiency. "What that does is it takes that attenuated Cas9 to a specific address in the genome. Then, because cas9 is tethered locally, it's at a high effective concentration." This gets around the pesky problem of off-target effects. The programmable DNA-binding domain adds a third layer of security that Cas9 will only cut where you want it to, because it can only stick onto DNA when tethered to the other protein.
In the study, the scientist performed an unbiased, deep-sequencing method to detect off-target activity. The protein fusion only had activity at three off-target sites compared to more than 50 with the wild-type Cas9.
The DNA-binding domain could be either a programmable zinc finger protein or a transcription activator-like effector (TALE), technologies that preceded CRISPR/Cas9 in the genome editing field.
The study comes out of a collaborative project with Jeremy Luban, also of the UMass Medical School, to excise HIV from the genome, funded by a grant from the NIH.
"We'll be working on developing these nucleases to efficiently and precisely excise HIV from the genome," Wolfe said. "We're interested in both developing nuclease that efficiently excises latent provirus from the genome as well as understand the structure of provirus when it's in reservoir cells."
Wolfe's lab has been working with zinc finger proteins for years, so using them in combination with Cas9 seemed like a natural fit. The idea for the new study came about by trying to replicate how zinc finger proteins work in genome editing. The zinc fingers must be fused to a nuclease, usually Fok1, Wolfe said.
"Unlike Fok1, which cuts indiscriminately, the nice thing about Cas9 is you will have much more control over nuclease activity, because of the protospacer adjacent motif as well as the complementarity of the guide sequence loaded into [Cas9]. It gives you an extra level of control over sites you're going to cleave in the genome."
The scientists also looked at TALEs. "One worked, the other one didn't," Wolfe said. But besides familiarity and efficiency, the study focused on zinc fingers for other reasons as well.
Because Wolfe would eventually like to use the nucleases in clinical studies, delivery into infected cells was a consideration. In addition, zinc fingers are more compact and thus easier to get into cells than TALEs are, he said.
"A zinc finger protein that recognizes a 12 base pair sequence is about 140 amino acids in length, whereas a TALE recognizing the same element is more than three times as large," he said. "If you're thinking about using those in viral delivery systems, more compact DNA-binding domains have advantages. Cas9 is already a very large protein and fitting it into a [viral vector] is already problematic."
So far, the study was only accomplished in cell culture, and Wolf said the lab hasn't yet looked at any delivery methods that would be use for in vivo approaches.
Separately, Wolfe is also trying to use CRISPR/Cas9 to create a better model of HIV infection in primary cells, using gene editing to do targeted integration of HIV into the genome.
"When HIV normally infects a cell, it is in a semi-random position in the genome, [and] usually integrates around active genes," he said. "We'd like to be able to build cell models that have HIV in a single site in the genome. To do this in primary cells has been challenging. In transformed cell lines, you can create HIV integrants that you can define locations and expand cell locations, but for primary cells, this hasn't been done."
Transformed cell lines are good models but not necessarily perfect models of what latent HIV looks like in resting memory T cells, he said.
Wolfe has already shifted to making the Cas9/zinc finger fusions more specific to editing HIV. "The linker could be optimized further and probably result in improved precision," he said. "We'll be looking at doing that more when we optimize for specific targets sequences, like we will when we look at target sites in HIV."
He said that he has applied for a provisional patent on the technology but is unsure whether it will be acknowledged, given the general uncertainty over CRISPR intellectual property.
"We'd love to have someone license it, but we're not in any discussions right now," he said.