CRISPR/Cas9 complications

This subpage constitutes the fifth part of the theory for Biotech Academy’s material on CRISPR-Cas9.

Off-target effects and possible solutions

The first problem encountered with genetic modification with Cas9 is off-target effects. These are seen when the Cas9 protein interacts with a different sequence than the one planned to modify. It is clear that this can create major problems, as, for example, important genes can be destroyed unconsciously. The problem arises mainly because the sequence that the given sgRNA recognizes exists in several places in the genome, or that there are sequences that are sufficiently similar to it. The alternate PAM sequences 5′-NAG-‘3 and 5′-NGA-3’ may also be recognized to a lesser extent, which lowers specificity. Off-target effects are difficult to predict for individual sequences in a genome, which increases uncertainty when using Cas9. You can reduce the amount of off-target effects by trying to make sure that your sgRNA is as unique as possible in relation to one sequence in the genome. With the specificity from the PAM sequence and the 20 recognizing nucleotides, it is easy to achieve in organisms with small genomes, but can be much more complicated in complex organisms with large genomes, such as humans.

Cas9-nickase (Cas9n) is an alternative modified version of the Cas9 protein that can significantly reduce off-target effects. Here, either the RuvC or HNH nuclease is deactivated, by point mutations in the coding DNA sequence of the Cas9 gene. This means that the Cas9 nickas do not make double strand breaks, but instead so-called nicks, which are only breaks in one DNA strand. The mechanism of operation is exactly the same as the natural form of Cas9, except that only a single strand of DNA is cut. The point is that for the formation of a DNA fracture, you must use two Cas9 nickases, each of which must recognize each of the two DNA strands with different sgRNA. Thus, twice as much identification of DNA sequences compared to the use of natural Cas9 is required to achieve the same result. This results in an increase in precision. If the Cas9 nick gas mistakenly recognizes somewhere else, only one nick will be made in one DNA strand, which will be repaired much more accurately than a double-strand fracture. Double strand fractures formed by two nicks will have loose ends, but can still be exploited for genetic modification. See Figure 11. This approach increases the accuracy of recognition as more identifications are required while lowering the importance of off-target effects, as off-target nicks are not as severe as off-target double-strand fractures. In an experiment with the use of Cas9-nickase, specificity was increased 1500-fold compared to natural Cas9.

Figure 11. The Cas9 nickase (Cas9n) has only one functioning endonuclease domain, HNH or RuvC, and can therefore be used to make nicks in single DNA strands. In this way, you can use two Cas9n to make one double-strand break, thus ensuring a higher precision, as two different sgRNA, A and B, binding in approximately the same place, are needed. Thus, more correct sequence identification must take place before a double-strand break is achieved.

SpCas9-HF1 (Streptococcus pyogenes Cas9 High Fidelity 1) is a variant of Cas9 that lowers off-target effects, but preserves the effectiveness of genetic modification in the selected DNA sequences that you have designed your sgRNA to target. For SpCas9-HF1, all 20 nucleotides in the recognition sequence matter more and mismatches are not accepted, rather than it is primarily the first nucleotides after the PAM sequence that matter most in the normal type Cas9. Thus, a better match is required before SpCas9-HF1 binds, thus avoiding off-target effects due to lack of specificity. SpCas9-HF1 is formed by replacing 4 specific amino acids in Cas9 by mutations.

Cas9n and SpCas9-HF1 are good examples of how to manipulate the protein structure of Cas9 to improve the precision of gene modifications.

Another promising approach to avoid these unfavorable off-target effects is through bioinformatics, where computer algorithms and data processing can predict which sgRNA sequences give the best results. This computerized approach to genetics allows processing large amounts of genetic information, and already plays a major role in modern genetic engineering. There are up to several online algorithms that try to predict off-target effects based on genetic data.

NHEJ lowers the effectiveness of homology directed repair

In eukaryotes, like fungal cells and animal cells, NHEJ is a dominant repair mechanism in the treatment of a double-strand fracture. This means that the effectiveness of insertion of DNA sequences is limited by genetic modification, as HDR is suppressed. In order to achieve a higher efficiency of genetic modification with sequence insertions, NHEJ can be temporarily inhibited so that HDR can be used. There are various methods to temporarily inhibit NHEJ. This is achieved by inhibiting the activity of the enzymes that carry out the repair. When NHEJ is inhibited, the gene modification should take place, as HDR has its greatest effectiveness here. Subsequently, the NHEJ repair system should be reactivated, as a permanently blocked NHEJ system weakens the organism’s natural defenses against DNA damage, which is not favorable.
A team of researchers succeeded in increasing the effectiveness of HDR 8-fold in human cells by temporarily inhibiting NHEJ, confirming that it is an important factor to take into account.