CRISPR Editing, Then and Now
We’ve all heard the story of CRISPR/Cas9, the gene-editing giant. CRISPR is still talked about like a risky new technology, but since it was first harnessed in 2012 CRISPR has changed a lot.
Originally, CRISPR/Cas9 was able to make one change and one change only: cut. At a specific point on the genome, a Cas9 enzyme with a supplied sequence could break the DNA, forcing the cell to repair it ham-handedly with a process called non-homologous end joining (NHEJ). It was much better than the techniques used at the time, but it was still too unreliable, complex, and unstudied for widespread editing.
Now, a veritable cornucopia of modifications to the CRISPR/Cas system has made it cheap, reliable, versatile, and well documented. Specific bases can be replaced, new genes can be inserted, and complexity decrease has made it incredibly easy to use. All this, and its accuracy is already leaps and bounds above what it was only a few years ago, and always improving.
CRISPR’s Origins
To understand CRISPR’s improvements over the years we need to know where its functionality comes from.
In short, CRISPR/Cas is found in nature as a bacterial immune system. Just like humans, bacteria can be infected by viruses. The mechanism viruses use to infect bacteria is complex, but in order to understand CRISPR all we need to know is that one key step is the injection of viral genetic material into the bacterium.
The CRISPR half of CRISPR/Cas is a portion of the bacterial genome that stores copies of short (~20 base pair) sections of viral genetic material from previous infections. These copies can then be transcribed into RNA guide sequences, and used by a Cas enzyme to check for and stop the beginnings of a viral infection.
The Cas9 protein specifically does this by comparing the guide RNA to any DNA it finds and inducing a break if the two match up, deactivating the viral DNA and preventing the virus from moving to the next stage in its life cycle.
From here, it seems pretty obvious that you could make a break anywhere you want by putting in a guide RNA with the same sequence as the locus you want to mutate. The problems reveal themselves deeper within the functioning of the CRISPR/Cas immune system.
Problems and Solutions
There are four big problems with using CRISPR/Cas for editing, each with their own solutions: overall complexity, specific targeting, the types of mutation that can be induced, and off-target editing.
Complexity
One of the first barriers to be overcome in CRISPR editing was the complexity of the system. The problem is that the natural variant of Cas9 required multiple RNA complexes to bind together before it could be combined with the protein and scan for the complementary (matching) sequence. These are crRNA (CRISPR RNA), which contains the guide, and tracrRNA (trans-activating crRNA), which secures both RNAs to the Cas9 protein. These RNAs anneal (bind) to one another to form gRNA and after minor modification are finally used to guide Cas9 to the correct sequence.
It was soon discovered that with relatively little modification, one can just append the tracrRNA to the crRNA to form one continuous “sgRNA” (single-guide RNA), and it will work effectively the same. This meant that instead of needing to get two RNA sequences into a cell and then get them to assemble correctly, you only need to get one RNA sequence in, and it will fold on its own into the correct shape.
Existing techniques meant that everything could already be delivered to cells joined together in plasmid vectors, which basically means that the gene for the Cas9 protein and a gene for the sgRNA could be bundled together.
Specific Targeting
In the previous diagrams, you may have noticed a very short section of the target DNA labeled “PAM.” This stands for Protospacer Adjacent Motif, and it is one big limiter to where Cas9 can induce mutations.
The purpose of the PAM is to prevent bacteria from targeting their own genomes with the CRISPR/Cas immune system. Because the guide sequence is copied from a template stored in the bacterium’s DNA, the problem arises for the organism that the Cas9 will cut its own template. This is solved by forcing the Cas9 to only cut if a specific, preset sequence is present in addition to the sequence set by the guide RNA. Templates are only taken from virus genomes next to this sequence, but the short motif itself is not incorporated into the CRISPR array in the bacterial genome. This motif is the PAM.
This is great for bacteria because it lets them keep a precise record of the virus, inactivate viral DNA when it’s present, and not disturb the copies they keep. However, it limits the areas where Cas9 can edit to those next to PAM sequences, as the variable guide sequence has no control over the requisite PAM.
Specifically, the PAM for the variant of Cas9 from S. pyogenes is the sequence NGG, meaning Any Nucleotide, Guanine, Guanine. Cas9 cuts 3bp upstream from the PAM, within the guide sequence. For most targets, this NGG motif can be found close enough to the locus you want to mutate, but for many, it’s simply impossible.
This problem, unfortunately, doesn’t currently have a perfect fix. The best we have so far is the ability to use different variants of Cas9 with different PAM sequences, allowing for more targets to be reached. For current editing techniques, being able to edit a specific nucleotide is not usually needed, and as such mutagenesis within a reasonable margin of the intended locus is good enough. However, this is certainly not true for all applications of CRISPR/Cas editing.
Mutation Type
Perhaps the greatest limitation of the original Cas9 was its confined mutagenic repertoire. Basic Cas9 can only induce double-stranded breaks in DNA. The standard repair process for DSBs (NHEJ) is prone to errors, usually resulting in indel mutations that effectively ruin any section of the gene that comes after them. This is very useful for knockout mutations, or deactivating genes, as any DSB in the promoter of a gene would prevent it from being expressed. Much like mangling the lock on a door, it makes it such that even if you have the right key you can’t use it to get in.
But if you only want to change one nucleotide, or insert a whole new gene (a knockin mutation), you were out of luck.
Now, new developments are making both of these possibilities. For a long time, there has existed a method for inducing a knockin, but it was horrifically unreliable. In essence, it relied on a type of DNA repair called HDR, wherein a template that contained the DNA you want to add was also transported to the cell, and it would hopefully be used to repair the DSB induced by Cas9. This method hinged on long homologous regions on the ends of the template, such that they would match up with the DNA on either side of the induced break and initiate the alternative repair pathway.
The thing is, this only worked around 20% of the time, seriously hindering its functionality. New developments harness another natural pathway to drastically increase the frequency of successful insertions. Transposons are genes that have the ability to copy or excise themselves from their position in the genome and “jump” to another locus. Different transposons have different methods of locating the position they move to, and one category (Tn-7) has evolved to harness Cas proteins (Cas12k) which lack the ability to induce breaks, using them to mark insertion sites.
This method of creating knockin mutations with CRISPR/Cas is showing significant improvement over the old one, boasting close to 80% success rates.
For a long time, the realm of single nucleotide editing seemed impossible to access with CRISPR/Cas. But in recent years, we’ve seen it begin to come to fruition. This technology is still not fully fleshed out, but its results thus far are impressive, to say the least.
Converting one nucleotide to another is no easy task, as it is extraordinarily hard to edit the desired base without affecting the surrounding DNA. The new solution actually converts the adenine nucleotide into inosine, which has a bonding pattern that resembles that of guanine. The enzyme which does this doesn’t replace adenine with inosine, but actually rearranges the atoms of one to form the other.
To fully change the A-T pair to a G-C pair, a variation of Cas9 called Cas9 nickase is used to induce a single-stranded break opposite the new inosine, and a complementary cytosine is used to repair the nick. The ability to convert A to G complements the previously developed conversions from C to T and G to A. These used similar approaches with existing enzymes, but what makes A to G special is that an entirely new enzyme was developed to facilitate it.
While not all base change mutations are currently possible, the newly developed possibilities yield potential solutions to a multitude of diseases, and work continues to fully expand the available repertoire.
Off-Target Editing
The greatest complaint regarding CRISPR as a gene-editing tool has always been its accuracy. Off-target editing refers to when mutations are induced at positions on the genome other than the one specified by the guide sequence.
The clear issue with this is that it produces unwanted changes that could affect experimental results, and as such they must be screened for in test organisms and those which exhibit them filtered out of test groups. It also prevents CRISPR from being used in clinical settings, as too many unintended changes may go unnoticed and cause problems down the line.
There are a wide variety of ways to reduce off-target effects, each with their own benefits and weaknesses. Depending on the experiment, which of them are best to implement can change wildly, and as such we will focus on the most broadly applicable solutions.
Cas9 Nickase
As mentioned, Cas9 nickase is a variant of Cas9 which only breaks the backbone on one side of the DNA helix, creating a SSB as opposed to a DSB as Cas9 usually does.
By using two nickase enzymes with target sequences on either side of the desired mutation, it effectively puts a second checkpoint on the mutation as both enzymes must induce a SSB in order to cause a mutation.
In addition, for certain applications, the nickase target loci can be offset from one another by some amount, creating overhangs on either side of the break. This can be useful when attempting to knockin a gene with HDR, as it makes it easier for homologous regions to bind to the DNA surrounding the break.
Of course, the multiple guide sequences do again pose the problem of complexity, which generally makes many things more difficult and less reliable in other manners.
Shorter Guide Sequences
While it may seem counterintuitive, shortening the length of the RNA template to less than 20bp can actually decrease off-target effects. The reason this can work is that it makes mismatches between the DNA and gRNA more detrimental to binding, making it much less likely for Cas9 to bind to a site other than the target site. In fact, this has been shown to reduce off-target mutagenesis by 5000 times!
Altered Cas9
Many variations of Cas9 have been developed, from HypaCas9 to SpCas9-HF1. These versions of Cas9 are modified to generally make the binding of the DNA, sgRNA, and Cas9 to one another weaker. Specifically, the R-loop, where the DNA is unwound and checked for matching with the guide sequence, is often targeted by high fidelity Cas9 variants. Similar in principle to shorter guide sequences, disrupting the interaction of these three components, this time via modified protein, mandates that the guide sequence and DNA match more closely before cleavage can occur.
Additional Loop on sgRNA
This approach, like those with altered Cas9, seeks to destabilize the interactions between Cas9, the sgRNA, and the DNA. This time, however, it is done through a modification to the sgRNA sequence. By adding a hairpin motif to the 5' end of the sgRNA (end closest to the guide sequence). Because this is so close to the location where the R-loop forms, it physically gets in the way and makes R-loop formation more difficult. This decreases off-target effects by the same principle as the other approaches.
Overall, each solution to off-target editing presented here can reduce the effects by several orders of magnitude, and when used in conjunction they can produce incredible results.
Takeaways
- CRISPR/Cas has gained a wide variety of use cases since its discovery
- The biggest issues with CRISPR/Cas can be traced back to its origins as a bacterial immune system
- The greatest hurdles CRISPR/Cas has overcome are high complexity, low target specificity, a small mutagenic repertoire, and off-target effects
- Low target specificity due to the necessary PAM sequence and inability to induce all base change mutations are still ongoing problems in CRISPR/Cas editing
- Most solutions to off-target effects focus on making it more difficult to bind DNA to sgRNA to Cas9, so as to require a more accurate match of DNA to guide sequence before cleavage can occur
- A wide variety of other tech and biotech are being used in conjunction with CRISPR/Cas to solve these ongoing problems in gene editing