NEW YORK (GenomeWeb) – The discovery of bacterial CRISPR systems has created a cottage industry of researchers and biotechnology companies hoping to use the genome-editing method to cure disease or create better medications for a variety of illnesses. But there have been problems along the way — many naturally occurring CRISPR proteins are unsuitable for use in human cells, either because they don't work at all, work very ineffectively, or create off-target effects that end up causing more problems than they solve.
In a recent study published in Nature Communications, researchers from Merck KGaA life science business MilliporeSigma sought to address at least one of those challenges, that of CRISPR proteins being blocked from reaching human cells in order to edit them.
The team began by looking at the widely adopted type II-A Streptococcus pyogenes Cas9, SpCas9, in order to determine whether there was a way to eliminate its off-target effects, and possibly create a strategy for selectively editing identical genomic sites in different genes within the same genome.
"Unfortunately, many bacterial CRISPR systems either do not work or do not work well in human cells. Our team hypothesized that human chromatin proteins were blocking CRISPR biomolecules from reaching human DNA," Martha Rook, head of gene editing and novel modalities at MilliporeSigma, told GenomeWeb.
In exploring other CRISPR-Cas systems that may work more precisely than SpCas9, the team identified the type II-B FnCas9 from the Francisella novicida bacterium, noting that it possessed a novel enzymatic property that cleaves the target DNA in a staggered pattern, exhibiting a higher intrinsic specificity than SpCas9.
However, the researchers noted in their study, "FnCas9 is unable to cleave a large number of the targets that are efficiently cleaved by SpCas9 in human cells, even though the two Cas9 nucleases exhibit a similar activity on purified DNA substrates."
So, they developed a proximal CRISPR targeting method called proxy-CRISPR that restores FnCas9 nuclease activity in a target-specific manner. Importantly, the team added, the proxy-CRISPR strategy is applicable to a variety CRISPR-Cas systems, including type II-C Cas9 and type V Cpf1 systems.
"For unknown reasons, the bacterial SpCas9 protein works amazingly well in the chromatinized context of the human genome. Because of this, SpCas9 has become widely adopted, and we chose it to assume the role of 'hypothesized chromatin disruptor' in our initial proxy-CRISPR experiments," Rook said. "We made a mutant of SpCas9 that lacks DNA cutting activity (dead-SpCas9 or dSpCas9) and targeted it to adjacent sites (within ~50 bp) of other CRISPR systems (FnCpf1, FnCas9, CjCas9, etc.). To our surprise, proximal targeting of dSpCas9 had a strong and consistent positive effect on these various bacterial CRISPR systems in human cells."
And although dSpCas9 worked well to enhance other CRISPR systems, she added, proximal binding of dFnCas9 could enhance SpCas9 activity even further. "The idea is applicable to virtually any two CRISPR systems as long as the Cas9/Cpf1 proteins do not use each other's guide RNAs," Rook noted.
Importantly, proxy-CRISPR enhances the usability of native CRISPR proteins without the need to re-engineer them. Rook pointed to a version of the Cpf1 system called FnCpf1 that was shown to have the ability to edit the genome at locations that other proteins could not edit. Unfortunately, FnCpf1 was shown to have low activity in human cells. "[Our] proxy-CRISPR method was shown to rescue this low activity of FnCpf1," she said.
The researchers also hypothesized that proxy-CRISPR could be used to reduce the off-target effects of genome editing. "For a majority of targets, [DNA double-strand breaks] by an inactive CRISPR–Cas nuclease will require at least two guide RNA binding sites proximal to each other on a chromosome," the authors wrote. "The likelihood of two similar genomic sites occurring elsewhere in the genome greatly diminishes compared to one site."
They tested this hypothesis by looking for off-target sites corresponding to the editing of POR and EMX1 by FnCas9 with the assistance of dSpCas9 proximal binding. They found no similar genomic sites corresponding to the dSpCas9 proximal binding sites in the POR and EMX1 loci and no off-target cleavage by FnCas9.
They further tested their theory by using CjCas9 in combination with dSpCas9 to selectively edit two identical targets in the human haemoglobin subunit beta (HBB) and subunit delta (HBD) loci. "When expressed alone, CjCas9 was unable to cleave either the HBB target or the HBD target in K562 cells," the authors wrote. "But when expressed in combination with SpdCas9 and a pair of sgRNAs specific to two proximal binding sites in HBB, CjCas9 cleaved the HBB target efficiently (32 percent indels) without cleaving the identical target in HBD. Conversely, CjCas9 was able to selectively cleave the HBD target fairly efficiently (14 percent indels) without cleaving the identical target in HBB when it was aided by SpdCas9 and a pair of sgRNAs specific to two proximal binding sites in HBD." Here again, they found no off-target cleavage by CjCas9.
Having demonstrated its utility, MilliporeSigma has filed several patent applications on the proxy-CRISPR technology, Rook said. The company believes proxy-CRISPR can be used in a variety of genome editing projects to boost both gene knock-out and knock-in genome editing, enable targeting of DNA cutting closer to desired mutation sites, and reduce the frequency of off-target effects by creating a requirement for two binding events to achieve knock-out or knock-in.
"MilliporeSigma's proxy-CRISPR method will certainly provide new genome editing possibilities for our customers, business partners, and our internal cell engineering R&D teams," Rook noted. "We have a number of collaborative opportunities we will pursue to maximize the benefit of proxy-CRISPR in research applications that overcome technical hurdles for our customers."
The company is also exploring ways proxy-CRISPR can be used to enhance its own drug development programs, she added. "Our intention is to make this technology available to scientific research communities in academia, biotech, and pharma the same way we have developed and distributed TargeTron, ZFN, and CRISPR reagents over the last 12 years. For applications beyond research, we are open to licensing discussions," Rook said.