Two papers, one published today in Nature and the other last week in Science, describe a new mechanism in which bacterial transposons associate with CRISPR-Cas systems to allow for RNA-guided DNA insertion. Both papers show that a stolen CRISPR-Cas system can use its own guide RNA (gRNA) to recognize protospacers and, instead of cutting them by making a double-strand break, transpose adjacent to them.

Today’s paper, “Transposon-encoded CRISPR-Cas systems direct RNA-guided DNA integration,” is from a team of graduate students in the lab of Samuel Sternberg, PhD, an assistant professor at Columbia University in the department of biochemistry and molecular biophysics. In it, they show that the natural combination of a CRISPR–Cas and transposition systems from Vibrio cholerae can efficiently insert large fragments of DNA without making double-strand breaks.

“They investigated the biology of how the RNA-guided DNA targeting Cascade complex (Type I CRISPR) interacts with the transposition system (a protein called TniQ) in integrating the DNA, coupling the DNA targeting and inserting pathways, and harnessed this mechanism as a gene editing tool,” notes Gaétan Burgio, MD, PhD, group leader in the department of immunology and infectious disease at the Australian National University.

Gaetan Burgio, MD, PhD, The Australian National University [Multimedia, JCSMR, ANU]
Last week’s paper, “RNA-guided DNA insertion with CRISPR-associated transposases,” from the group of Feng Zhang, PhD, at MIT and the Broad Institute, showed, well, basically the same thing. They characterized a functionally similar system (albeit distantly related) from the cyanobacteria Scytonema hofmanni and identified the CRISPR-associated transposase, which they called CAST (or ShCAST, in this case.)


There are limitations associated with CRISPR-Cas9 DNA insertion which RNA-guided integrases may allow researchers to circumvent. When the Cas9 enzyme makes a DNA double-strand break, the cell undergoes numerous biochemical reactions that ultimately result in different outcomes, one of which might be the insertion event of interest. “By avoiding double-strand breaks,” notes Sternberg, “we avoid requirements on those host factors, as well as higher purity of the editing outcome, elimination of the DNA damage response, and lower risk of genomic deletions or translocations.”

The two papers have multiple similarities not the least of which is that they both prove a hypothesis laid out two years ago by Joe Peters, PhD, professor in the department of microbiology at Cornell University and Eugene Koonin, PhD, senior investigator in the evolutionary genomics research group at the National Center for Biotechnology Information (NCBI).

Connecting the dots

The idea that transposons and CRISPR are linked has jumped around for a while. Sternberg tells GEN that “a 2017 paper from Joe Peters and Eugene Koonin was the first to connect the dots.”

Joseph Peters, PhD, professor, Cornell University

The 2017 PNAS paper, “Recruitment of CRISPR-Cas systems by Tn7-like transposons,” was the first report of Tn7-like transposons that contain minimal CRISPR-Cas systems. The researchers identified three distinct groups of Tn7-like transposons that encode minimal variants of type I CRISPR-Cas systems which, they write, “implies that they are competent for pre-CRISPR RNA (precrRNA) processing yielding mature crRNAs and target binding but not target cleavage that is required for interference.” In fact, they suggest that, “these transposable elements are predicted to propagate via RNA-guided transposition, a mechanism that has not been previously described for DNA transposons.” A fourth system, that was independently acquired by Tn7-like transposons, was found later by the same groups—this time a type V CRISPR-Cas system.

Indeed, Peters tells GEN that “we found examples where the protospacer that matched the gRNA encoded in the element was immediately adjacent to the element suggesting it was used to recognize the target. But, we didn’t see it actively in the lab.” The Sternberg and Zhang labs were able to accomplish this by expressing the transposition proteins in E. coli, along with the CRISPR-Cas system and the gRNAs of their choosing, to see the insertions occurring in the fashion that they predicted.

Transposons are mobile elements that move around the genome. Although many jump around aimlessly without selecting a target, some are more choosy. Tn7—the transposon at the center of both papers—is in the more discerning class. It has a core machinery (consisting of ABC proteins) that can work with one or two other proteins (D and E) to choose different, and very specific, targets. They use the D protein to jump into a safe site in the chromosome and E to jump into actively replicating conjugal plasmids to hitch a ride to another host.

One of the early observations that seemed interesting to Peters was that the Tn7 elements that had picked up the variants of the CRISPR system still had the ABCD components and could go to the safe sites in the chromosome. But, the E protein was replaced with a variant of the CRISPR-Cas system that could recognize a protospacer but had no ability to cut it. The fact that happened four different times over evolution indicates that there was something about these elements that primed them to pick up new, cool things. That was “the first clue,” notes Peters, “that they might be using this for horizontal transfer.”

The finer points

“When we set out to interrogate the specificity of our integration system—how accurate it is in integrating our genetic payload at just one genomic target site—we were thrilled to see that the Vibrio cholerae system is extremely high-fidelity, with an on-target accuracy >95% across dozens of distinct target sites,” notes Sternberg. He adds that “this is an interesting point of distinction with the CRISPR-associated transposase described by Zhang and colleagues, which exhibits a relatively low degree of accuracy and promiscuosly inserts DNA at many off-target sites.”

“We programmed our system to move to a different site dozens of times, and it worked every single time,” notes Sanne Klompe, first author on the paper and a graduate student in the Sternberg lab. She adds that “the high degree of specificity together with its programmability makes this system ideal for genome engineering applications.”

This system may, indeed, be an ideal choice when inserting large amounts of DNA. The transposon has evolved to insert DNA without double strand breaks, removing any worries about the host’s homologous recombination abilities. Peters explains that, “you can, for example, insert a good copy of a very large piece of DNA or even a series of genes into a place.” This holds great promise because “you can put it in anywhere you want, very efficiently.” However, he notes that where this system does not make a lot of sense to use is in making very small, precise DNA changes. “For that” he adds, “double-strand breaks would make more sense.”

The big question

Researchers will, undoubtedly, put this system to good use in their labs. But, for the rest of the world, there is a looming question. Will this work in a mammalian system? Burgio skeptically notes that, “while gene replacement in eukaryotes was touched on in the Zhang paper, I strongly suspect it might not work as well as it does in bacteria for a few possible reasons.”

“The system is quite large to deliver into the cells. With three genes, a donor vector and the cascade complex, which is composed of three genes, it is significantly larger than the classical Cas9.” In order to work, these genes would have go into the nucleus, transcribe, translate, and assemble in a short period of time. Although Burgio notes that one possibility could be to deliver it as a protein complex, “it would be quite massive” and he doubts that he would be able to add the complex into a micro injection needle and deliver this into a mouse embryo. Additionally, because the transposition system is derived from bacteria, he notes that “we have no idea how efficient it will be in mammalian cells.” And, there could be scars left after transposition.

Lastly, he notes that this simply may not be the technology that is left standing at the end of the day as there are other, existing systems such as homology-independent targeted insertion (HITI) that show a lot of promise in mammalian cells to insert large chunks of DNA while bypassing non-homologous end joining (NHEJ)—one pathway that repairs double-strand breaks in DNA.

The four authors of the paper, all members of the Sternberg lab at Columbia University: Tyler Halpin-Healy, Sanne Klompe, Sam Sternberg, and Leo Vo.

These two papers, together with the groundwork set in 2017, have opened the door to a new paradigm of RNA-guided, DNA targeting and inserting. The best part of publishing this work, “by far” according to Sternberg, “was the opportunity to work with three extremely talented, motivated, and fantastic PhD students.” And, there is likely a long road ahead. “We screened a number of bacterial transposons, and so far, the transposon from V. cholerae has performed the best. There are hundreds, probably thousands, of other transposon variants in nature that have also repurposed CRISPR–Cas systems for RNA-guided DNA insertion, and so there is certainly still more to be discovered,” notes Sternberg.

Previous articleGenetic Makeup Influences Immunity to Childhood Vaccines
Next articleCelyad Faces “Go-or-No-Go” Decisions on Lead CAR-T Candidates