In a recent academic collaboration, scientists discovered another surprising behavior of the Cas9 and Cpf1 enzymes – a behavior which may have implications when developing novel therapies with CRISPR. We spoke with Brett Robb, the Scientific Director for RNA and Genome Editing at New England Biolabs Inc, and Becky Fu (previously a graduate student at Stanford University and now a Postdoctoral fellow at UCSF) to learn more about their research published in Nature Microbiology.
Molly Campbell (MC): Please can you explain the terms "nicking" and "full cleavage" in the context of CRISPR?
Becky Fu (BF): CRISPR nucleases are RNA-guided proteins that have the ability to cleave double stranded DNA. CRISPR nucleases have been characterized as either making blunt or staggered DNA ends after cleavage of both strands, depending on the type of CRISPR nuclease in question. To add to the genome editing tool box, some CRISPR proteins have been modified through engineering to only cut one strand of DNA (a.k.a. nicking). This work (from Feng Zhang and George Church's labs in Boston) dovetails with work that Nancy Maizels and colleagues at University of Washington showing that nicks can be potent foci for genome editing.
MC: Why is it important to identify and understand potential off-target effects of CRISPR?
BF and Brett Robb (BR): It is important to identify and understand potential off-target effects of CRISPR-based DNA cleavage because this process has become a ubiquitous tool in research and is under intense development to produce potentially clinically relevant interventions. CRISPR tools are not 100% accurate and can potentially introduce unintended DNA changes. This is not just a problem for studying biological questions; as science pushes towards new therapies to move from the bench to bedside, it will be critically important to understand both the opportunities and limitations of the methodology.
MC: Can you please provide an overview of the methods and technologies you adopted in this study, and why you specifically chose them?
BF: We challenged plasmid libraries that contained mutations in target sequences with CRISPR protein programmed with a guide RNA. This method was used because the plasmid libraries contained thousands of mutations to four different targets which allowed us to access nuclease specificity in a high-throughput manner. Initial experiments unexpectedly showed mutations in certain regions of the target produced cleavage scores that differed with and without the linearization of the plasmid library after the exposure of CRISPR protein/guide RNA complexes. Looking into this in more detail, we realized that the plasmid library topology was actually illuminating differences between targets in the Cpf1 and Cas9 nicking and cleavage abilities.
MC: Your research suggests a previously unknown dual capability of CRISPR-Cas nucleases to initiate genetic change through nicking and full cleavage. What impact will this new knowledge have on the CRISPR field, and what doors does it open in terms of developing human therapeutics?
BF: The dual ability of CRISPR nucleases to nick and cleave DNA has implications for the use of CRISPR nucleases as genome editing tools and for basic biology of the bacterial immune system. Nicking can be induced with relatively little DNA and guide RNA similarity.
Experiments using CRISPR nucleases to perform genome editing may have a set of off-target considerations that involve nicking and that were not previously recognized. This is especially important for bench to bedside therapies. This work could inform new off-target prediction software or updates to existing algorithms that many researchers use to choose functional CRISPR nuclease guide RNAs with minimal off-target effects.
BR: How unintentional nicks in genomic DNA are repaired could have consequences for their use in experimental cells and organisms and for therapeutic use. Neither of us are involved in the development of human therapeutics, however, the off-target analyses for potential CRISPR nuclease therapies that we’ve seen reported are extensive and rely on multiple methods. A potential impact of this work on human therapies might be the careful consideration of nick repair as a potential source of off-target effects.
BF: In addition, the ability of the CRISPR nucleases to use little homology to induce nicking may imply a mechanism for which the CRISPR bacterial system has evolved. When invading genetic elements such as plasmids, transposons, or bacteriophages attack bacteria, the CRISPR system acts as an immune system. The CRISPR immune system takes in DNA from the invader and produces gRNAs to the invading element, this is known as spacer acquisition. Once the gRNAs are produced, CRISPR nucleases use them as guides to find and eliminate them via DNA cleavage. It is known that the CRISPR proteins Cas1, Cas2, and CRISPR nucleases work together for acquisition of bacterial immunity. And it has also been observed that CRISPR systems with existing spacers with little or no homology to the invading genetic element will allow for efficient spacer acquisition that leads to immunity. Our results suggest the ability to use lax homology to nick and anchor the CRISPR nucleases may aid in spacer acquisition.
MC: You offer companies access to a "guide-specific nicking repertoire". Please can you tell us more about this and why you think it is important for companies to access it?
BF: CRISPR nucleases are ubiquitously used in vitro and in vivo for genome editing in experiments, biotechnology, and potentially clinical applications. Understanding the full spectrum of off-target effects is critical for the safe use of CRISPR technology. Many studies have focused on the potential off-target cleavage events, but the fact that lax guide RNA homology to target can initiate DNA nicking was not known.
BR: Although nicks in the DNA in vivo are not considered as detrimental as double stranded breaks, DNA nicks can lead to damage repair that results in unwanted mutations. Therefore, understanding the requirements of nicking can help companies using CRISPR technology to select guide RNAs that avoid off-target effects and/or allow the identification of off-target nicking.
MC: There is great ethical debate surrounding CRISPR therapies. In your opinion, how do we approach the conversation of how CRISPR is applied in humans, and will your research have an impact this conversation?
BF & BR: The views that we share here are not necessarily representative of either Stanford University of New England Biolabs and are definitely not official positions of either organization. This being said, we believe that we should approach the application of CRISPR technology to human health as we would with any new technology.
There are multiple levels of regulatory bodies, from institutional review boards to national governmental agencies, that carefully weigh the risks vs medical need along the path of developing new human therapies. These bodies are composed of technical and regulatory experts and usually have input from ethicists. They are already in place and are well equipped to deal with CRISPR/Cas technology. All responsible researchers comply with existing regulations and the recommendations of these bodies. Doing otherwise is seriously detrimental not only to their own research, but risks harming the entire field.
Our findings won’t likely impact the larger conversation around the ethics of applying CRISPR-nuclease based technology to treating human disease. Experts may consider our results and factor them into their risk assessments as we discussed above.
Becky Fu and Brett Robb were speaking with Molly Campbell, Science Writer for Technology Networks.