Oh, it's not an easy job…
Human genes turned out to be difficult to fix
Due to the diversity of the human genome, the CRISPR/Cas editing system can edit a lot of excess
Kirill Stasevich, "Science and Life"
The genome editing method called CRISPR (or CRISPR/Cas), which appeared just a few years ago, quickly gained popularity among genetic engineering specialists. Initially, it was discovered as a bacterial antiviral protection system.
Bacteria keep pieces of viral genes in their genome (these sequences in the bacterial chromosome are called CRISPR), and when foreign DNA appears in the cell, special enzymes (Cas family proteins) compare it with viral samples – and if there is a similarity, the foreign DNA is cut and sent to scrap.
Biotechnologists at some point figured out how to use this system for their needs. Bacterial proteins, with the help of which they destroy the DNA of viruses, have been adapted to work in animal cells. In fact, the principle of operation remains the same: the protein looks for a site in the cell chromosomes that needs to be cut out, and as a "guide" the enzyme is given an RNA molecule with the same sequence of nucleotides as in the desired site. By comparing the RNA he carries with him with cellular DNA, the enzyme eventually finds the right place in the genome, and cuts it out. If there was a mutation here, it will disappear – the cellular DNA repair systems will plug the hole themselves so that there will be no mutation here.
Let's say again that this is a very simplified description of how the CRISPR genomic editor works. Now it already exists in several variants, with different proteins that cut DNA one way or another. But in any case, it is clear how great prospects are opening up here for bioengineering and medicine: a huge number of diseases develop due to defects in our DNA, so a tool that would allow such defects to be eliminated would be very useful. Of course, there were methods in genetic engineering before that that allowed editing DNA, but – and this is important – the CRISPR/Cas system greatly surpasses them in accuracy.
However, as for accuracy, it turned out that everything is not so simple – just recently we wrote that the CRISPR method introduces many unpredictable mutations into the genome. And now another article on the same topic has been published in Nature Medicine. David Scott and Feng Zhang, one of those who first came up with the idea of adapting bacterial CRISPR/Cas to biotechnological practice, from the Broad Institute came to the conclusion that the CRISPR editing method is not yet suitable for working with human genes, because human genes turned out to be too diverse.
The problem is that the same gene may differ in different people. As we know, DNA is a sequence of four different nucleotide molecules, four letters of the genetic code. For various reasons, one of the letters can change to another, without any harm to the gene and the cell. Such single-nucleotide substitutions occur constantly, and as a result, the same gene in two different people encodes a completely healthy, functional enzyme, but the sequence of the gene itself differs by one, two, three, several letters. And if we take a sufficiently large group of people and analyze their genomes for a gene, we will find a lot of variants in its sequence; there is a special name for this - single–nucleotide polymorphism.
And now let's remember that the genome editing system is looking for the place that needs to be edited using a special template (or guide) – an RNA molecule. It is synthesized so that its sequence coincides with the sequence in the DNA where the break needs to be made. But the template molecule is made quite short, it recognizes only a small piece of the edited gene. But the shorter the sequence of nucleotides, the more likely it is that exactly the same sequence will occur not only in the right gene, but also somewhere else. Given the wealth of single-nucleotide substitutions in the human genome, it seems quite likely that this is exactly what will happen: the CRISPR/Cas machine will sail to a completely foreign gene, which, due to the single-nucleotide substitution, has become similar to the one that was supposed to serve the real purpose.
For the first time, the fact that the diversity of genes can mislead CRISPR/Cas was talked about three years ago, but now it has been possible to quantify the scale of the problem. David Scott and Feng Zhang modeled RNA templates for twelve genes that are related to a range of diseases. Then, in order to understand what else these patterns can sit on, the authors of the work used full-genome sequences from genetic databases that collect information about the diversity of the human genome.
In some cases, the RNA template of the editing protein turned out to be really very accurate - that is, it bound only to the gene for which it was synthesized. But in other cases, the number of errors reached 10,000 – that's how many points in the genome, in addition to the right one, could be edited by a CRISPR/Cas machine. Moreover, in this case we are not talking about junk DNA, which does not encode anything, but specifically about genes encoding proteins.
But the CRISPR/Cas method still seems too convenient to just abandon it, and most likely biotechnologists will do everything possible to improve its accuracy. And here, firstly, it is possible to model the sequence of the RNA template more carefully; Scott and Zhang themselves proposed their own algorithm to improve the accuracy of the editing system.
Secondly, if you are going to apply genome editing to some patient, you should read his genome in full, and already with a genome-wide map on hand, knowing all his single-nucleotide substitutions, select the sequence for the RNA molecule that will lead the editing protein to the right gene. However, CRISPR/Cas will not be tested on humans soon – too much still needs to be tested on molecules, cells and animals.
Portal "Eternal youth" http://vechnayamolodost.ru
25.08.2017