Harnessing the processive nuclease Cas3 for large deletions

We have developed and demonstrated the utility of CRISPR-Cas3 for the targeted removal of large segments of DNA, from 20-400kb at a time- in Pseudomonas aeruginosa, Pseudomonas syringae, Escherichia coli, and Klebsiella pneumoniae. Written by Lina M. León on behalf of the CRISPR-Cas3 team

Like Comment
Read the Article

Bacteria have been historically useful model organisms for genetic studies, serving as a relatively simple platform to learn about the most basic mechanisms of life. However, even genetically tractable models like E. coli aren’t completely understood, with large sections of their chromosomes being ill-defined functionally, especially in mobile regions. And while serving as models for important cellular processes that bacteria have in common with other prokaryotes and eukaryotes has been fundamental, bacteria are interesting in their own right, hosting a diverse repertoire of metabolic enzymes, surviving in extreme environmental conditions, causing disease, and fending off competitors and predators. We aimed to develop a new tool for the molecular dissection of these abundant, infinitely interesting life forms.

Traditionally, analysis of bacterial chromosomes has involved methods like transposon mutagenesis and single gene knockouts, which have allowed us to assay phenotypic changes in response to gene perturbations. However, no tool has enabled scientists to selectively remove large segments of the bacterial chromosome at once. Such a tool would be particularly useful for probing bacteria, since bacterial genes are often grouped in functional units called operons, where genes involved in a specialized function are clustered together. For example, the lac operon clusters genes needed for lactose metabolism and expression of these genes is tightly regulated depending on available energy sources1. A more contemporary example is that of “immunity islands”, where bacteria encode distinct operons mediating phage defense in a single locus2. Additionally, bacterial chromosomes are packed with accessory regions, some of which are mobile, like prophages and conjugative islands3, often with no ascribed function. Since our group specializes in the biology of phages and CRISPR-Cas immunity in bacteria, we realized that there was untapped potential in the common Type I CRISPR-Cas systems for a tool that allows you to instantly remove groups of genes at once.

Type I CRISPR-Cas systems utilize a dual nuclease-helicase, Cas3, to target DNA elements that have infiltrated into the bacterial cell, distinct from the more common gene editing enzyme Cas9, which remains static on its target. We reasoned that directing Cas3 to a chromosomal region, and selecting for survivors, would allow us to isolate cells that had the targeted region removed . Pioneering work in this area had shown that this was possible with mobile regions4 but we wanted to assess how and if bacteria could recover from targeting events in non-mobile regions of the chromosome. To harness Cas3’s dual nuclease-helicase activities, we began by transplanting a CRISPR-Cas Type I-C system from an environmental isolate of Pseudomonas aeruginosa into the chromosome of a common laboratory strain, PAO1, for easier troubleshooting. We chose this subtype because both because it consists of only three component proteins for the crRNA-guided surveillance complex making it relatively streamlined compared to other Type I subtypes5 , but also because it is a generally understudied subtype.

In CRISPR-Cas systems, a repeat-spacer-repeat (R-S-R) unit is needed, together with the Cas proteins, to form a targeting complex, via the secondary structure of the “repeat” sequence, while the “spacer” directs the Cas enzymes to their target. This is accomplished when the RNA spacer hybridizes with the DNA target, followed by cleavage of the DNA by a Cas enzyme. Initially, providing a single R-S-R led to very low deletion efficiencies, It turned out that the selective pressure we were putting on the cells -having the CRISPR-Cas system create a lethal cut in the host chromosome- was leading us to select for survivors that had merely recombined the R-S-R so that the spacer was completely lost, therefore eliminating the risk of self-targeting entirely. To circumvent this problem, we mutated one of the repeat sequences (R-S-R*) so that it would still form the correct secondary structure but was no longer homologous to the other repeat sequence. Now that the cells could not homologously recombine the flanking repeats, our efficiency shot up to 90-100%. As the reader can imagine, this was a fantastic day in the lab! 

Using our highly efficient system, we turned our attention to a proof of concept experiment we were very excited about: using Cas3 to iteratively remove nonessential regions from the P. aeruginosa genome, regardless of the assigned function for genes in those regions.  Rather than build a minimal genome from the ground up, we instead asked whether we could reduce a chromosome  by isolating survivors of Cas3 self-targeting. The former approach has been successful in engineering a minimal Mycoplasma genitalium genome of only 473 genes, an exciting advance in synthetic biology6. Using our approach, we were able to remove nearly 1 Mb of the 6 Mb P. aeruginosa chromosome, demonstrating the utility of Cas3 for this purpose. We are excited about the engineering possibilities and imaginative applications of Cas3 for similar purposes. 

As a second proof of concept for the utility of using Cas3, we sought to engineer a plant pathogen, in collaboration with Dr. Jennifer Lewis and her graduate student, Ilea Chau-Ly, at the University of California, Berkeley. We chose the Pseudomonas syringae pv. tomato DC3000 strain, which encodes virulence effector genes in clusters, but importantly, does not have its own CRISPR-Cas system. Therefore, we cloned the Type I-C CRISPR-Cas cas genes, plus a spacer cloning site, to make an “all-in-one” vector that would allow us to quickly add different spacers depending on the site being targeted. This plasmid could then be used to transform a strain of interest.  We designed spacers to target Cas3 to the P. syringae virulence clusters and isolated survivors that had these elements deleted. These newly modified strains could then be subjected to in planta infection experiments to determine their phenotype. Our team envisions that this technique could be used to probe other pathogens that cluster virulence effectors into discrete chromosomal regions. 

We are not only excited by the prospect of harnessing our engineered system’s capabilities in non-model organisms, but also about the possibility of turning a strain’s own CRISPR-Cas system onto itself. Type I systems are ubiquitous in nature5, so your organism of choice may very well encode its own engineering toolkit. In addition to engineering the repeat to prevent recombination, and developing an approach to silence endogenous anti-CRISPR proteins, we also attempted to direct repair events to a specific outcome, which has been successful for many other groups making small deletions, by delivering a crRNA and a homology directed repair template. We were able to edit the environmental isolate that we cloned the Type I-C system from by simply delivering a spacer, and, our collaborator, Dr. Alenjandro Vasquez-Rifo at the University of Massachusetts Medical School, edited the PA14 Pseudomonas aeruginosa strain with its endogenous Type I-F CRISPR-Cas system. Moreover, with a DNA repair template provided, we were able to rein in the mighty Cas3 nuclease, allowing us to precisely define the boundaries of large deletions, which was inefficient or not possible when compared to Cas9.

Of course, many other challenges lie ahead. Differences in DNA repair mechanisms and transformation efficiency will need to be addressed for a given host. Additionally, we are excited to whether the ease of HDR for large deletions we observed can be harnessed for insertions of DNA cargo of various sizes. Our hope is that  the CRISPR-Cas3 system developed here (PaeCas3c) will be a useful addition to the geneticist’s toolkit as we come to appreciate the diversity and utility of microbial life.

Read our full study: https://www.nature.com/articles/s41592-020-00980-w


  1. Jacob, F. & Monod, J. Genetic regulatory mechanisms in the synthesis of proteins. J. Mol. Biol. 3, 318–356 (1961).
  2. Doron, S. et al. Systematic discovery of antiphage defense systems in the microbial pangenome. Science 359, eaar4120 (2018).
  3. Brockhurst, M. A. et al. The Ecology and Evolution of Pangenomes. Curr. Biol. 29, R1094–R1103 (2019).
  4. Vercoe, R. B. et al. Cytotoxic Chromosomal Targeting by CRISPR/Cas Systems Can Reshape Bacterial Genomes and Expel or Remodel Pathogenicity Islands. PLoS Genet. 9, e1003454 (2013).
  5. Makarova, K. S. et al. Evolutionary classification of CRISPR–Cas systems: a burst of class 2 and derived variants. Nat. Rev. Microbiol. 18, 67–83 (2020).
  6. Hutchison, C. A. et al. Design and synthesis of a minimal bacterial genome. Science 351, aad6253–aad6253 (2016).

Lina Leon

Graduate student, UCSF

1 Contributions
0 Following

No comments yet.