Skip to main content
Premium Trial:

Request an Annual Quote

NHGRI Scientists Create Pipeline for Multiplex CRISPR/Cas9 Gene Editing


NEW YORK (GenomeWeb) – Scientists have developed a method for generating large numbers of mutated genes with CRISPR/Cas9 and a pipeline for identifying interesting mutations that enables high-throughput, multiplexed gene editing research. They published their work earlier this month in Genome Research.

"What this paper did was lay out a pipeline for making mutants at a relatively high throughput," senior author Shawn Burgess of the Translational and Functional Genomics Branch of the National Human Genome Research Institute told GenomeWeb. "Almost anything can be done on a small scale, but everything changes when you try to increase numbers."

The researchers were able to target zebrafish genes using CRISPR/Cas9, achieving germline mutations in about a quarter of the fish in the study. They targeted 162 locations in 83 genes, successfully mutating 82 of the 83 genes in the study, while using only about 1,000 fish. Burgess added that scientists in his lab have created about five times as many mutations in zebrafish with CRISPR/Cas9, but just haven't published data on them yet.

Creating mutations with CRISPR/Cas9 was six times more effective than other gene editing technologies, including both zinc finger nucleases and transcription activator-like effector nucleases, Burgess said, even when those other methods were optimized. And when comparing CRISPR/Cas9 to pre-gene editing technologies, like random mutagenesis followed by exome re-sequencing, it's not even close.

"You would have to go through usually five to ten thousand genomes to find the mutant you wanted," Burgess said. To find just one mutation, the cost of reagents alone could be $20,000 to $30,000. "Now the reagent cost for this is $30," he said.

Multiplex gene editing had been demonstrated in mouse and zebrafish before, Burgess said, but this paper showed it was possible to do it in a large-scale, repetitious fashion.

"You can get multiple mutations in the same fish with very little loss in efficiency compared to doing each one alone," he said.

"In a shorter time, I can target many genes," Co-lead author Gaurav Varshney of the NHGRI told GenomeWeb. "Instead of doing 100 individual injections, I can do 10 injections. Instead of doing 100 out-crossings, I can do 10."

And soon scientists could create triple mutants, targeting three potentially related genes at the same time. Doing such studies in mice is difficult because of the mathematics of animal husbandry, Burgess said, since only about one in 64 will have all three genes knocked out and mouse litters are only about eight to 12 pups. "Zebrafish can generate several hundred embryos, so you can get a number of [mutants] just by doing brute force genetics," he said. "We're moving in that direction now, and it's actually working in our hands."

Creating that many mutants meant that the scientists had to develop a way to find all those mutations, which they did with a combination PCR and sequencing workflow, which Burgess said could be applied to any model organism, not just zebrafish.

The first step was to perform amplification of the target DNA. That PCR product was split in two and half labeled with a fluorescent marker, half with a sequence barcode. The scientists put the fluorescently marked DNA into a capillary sequencer to quickly find insertions and deletions. "We used capillary sequencing to save fish we think are important," Burgess explained.

The researchers then took the DNA from those interesting fish and sequenced it on an Illumina MiSeq instrument, revealing the precise sequence of the mutations spotted earlier. This split sequencing approach allowed the mutation identification process to work smoothly and efficiently, Burgess said.

Genotyping the fish this way also prevented the scientists from accumulating thousands of fish sitting around in tanks and kept the animal husbandry aspect of the project from bogging it down.

The authors also reported several findings about the mechanics of CRISPR/Cas9 genome editing.

One way to make guide RNAs is to synthesize them with the T7 polymerase, which requires the first two bases to be "GG," Burgess said. A previous paper had reported that simply adding "GG" didn't hurt the efficacy of the guide, but Burgess said they found evidence against that. "Having those extra Gs hanging off the end has a strong negative impact on success for the guide," Burgess said, adding that it was three times more likely to fail if there were mismatches at the 5' end of the guide.

He also said they found very little impact from off-target activity. "We focused on off-targets that would have mutated other genes, as opposed to off-targets between genes, and we saw almost none," he said. "For the kind of work we do it doesn't matter so much anyway because we can out cross, in breed, and look at linkage to the mutation. For us it's not that important an issue, and even so we weren't seeing a lot of problems."

Varshney said that his next step was to investigate over 100 genes that humans share with zebrafish that are predicted to be associated with deafness. In a study of seven genes, he found that mutations in three of them recapitulated symptoms of human deafness in the fish. Studying that many genes used to take more resources, but now that study can be done more efficiently.

"Something that would have taken five people can now take two people to do the work," he said.