Genome Modification—a Practical Approach

http://blog.allelebiotech.com/2013/07/talengenome_editingcrisprcas/

The ability to modify genomes has always been fervidly sought after by molecular, developmental biologists and geneticists as it would provide them with the means for finding out what a particular piece of the genome may do in the biological process they are studying. The discovery of naturally existing P-element helped a generation of Drosophila geneticists and made the fruit fly a prime model system for gene function studies in the 80’s and 90’s. But P-elements inserted at uncontrolled sites, making it essentially a gene transfer vehicle without much control. The introduction of prokaryotic recombination systems, e.g. LoxP and Cre, provided researchers with tools to obtain more control of the inserted genes in a host chromosome during a biological process such as development. Transposons like Sleeping Beauty, Piggybac, or Tol2 made similar experiments possible in mammalian cells.

Still, the randomness of transposon-type elements’ insertion, much like retrovirus or lentivirus, could cause trouble if they land in an undesirable spot. Methods of inserting transgenes only in well-known, harmless, and transcriptionally active regions, so called “safe harbors”, were subjects of interest of researchers and NIH grant topics in the past couple of years under “directed genome editing”. Gene knock-out or knock-in can be achieved through vector-mediated homologous recombination such as the rAAV genome engineering system and the “TARGATT” system, which are commercially available as kits or services.

However, instead of inserting an exogenous gene, it is often highly desirable to modify an endogenous genome sequence, which requires the modification apparatus to first recognize the target sequence. ZFN and TALEN both recognize DNA targets through specific nucleotide binding protein domains, with TALEN having more flexibility if assembled in a “Lego”-like format because each domain can specifically recognize a “C”, “G”, “T”, or “A” base. The description of using CRISPR/cas system in a recent burst of publications opened up new ways of binding to specific DNA sequences and nicking or severing the dsDNA. This system does not require engineeredDNA binding domain assembly; instead, it uses a guide RNA to find the target DNA sequence to direct endonuclease, in a sense quite like RNAi. However, the enthusiasm about CRISPR/cas was somewhat dampened by a report last month in Nature Biotechnology that reported off-target effects of CRISPR/cas was much higher than ZFN and TALEN. Particularly, if mismatches are located in the 5’ portion of the guide RNA targeting sequence, they can be well tolerated up to 3 or 4, even 5 mismatches. Unfortunately this is also similar to the tolerance of the RNAi matching region outside the core 12-base region. The difference is: for RNAi, the off-target damage is temporary and ignorable if the extent is insignificant compared to the effects on the intended target while for CRISPR/cas, an off-target cut on the chromosome is permanent.

On the positive side, in an even more recent publication in Nature Methods, mutant strains of C. elegans were obtained using the CRISPR/cas system and no evidence was found for off-target changes, at least not in an overwhelming fashion. Much value of the estimates of off-target effects relates to the methods used for analysis. Currently, most of the studies looked at potential off-target sides by searching for partial matches. In the future, whole genome sequencing will be increasingly required for submitting such publications.

On a practical note, if you intend to take a dive and try to use any one of these methods, your number one problem will be that none of the methods will result in 100% modification even if you can ignore the off-target problems for now. Therefore, many of our customers ask about a screening strategy. One could use traditional drug selection and fluorescent protein (FP)-based sorting, but these can only help you find cells that are successfully transfected with the ZFN, TALEN, or CRISPR/cas expressing DNA molecules, not necessarily having the genome modification result. We have formulated the idea of inserting the target site into an FP-bearing plasmid as a surrogate target cutting indicator, and use another FP to track transfection of the TALEN plasmid. Nonetheless, in the end, PCR-amplifying the target region of the chromosome and doing either an enzymatic mismatch detection assay (e.g. T7 endonuclease) or sequencing is the only way to know for sure whether genome editing has occurred.