Scientist Stories: Martin Jínek, CRISPR/Cas Genome Editors from Bacteria to Biotech

Hello everyone, welcome to another TCSS talk. This talk will be about an hour long and afterwards there will be a Q&A session. Both will be live streamed on YouTube and also recorded. If you have any questions, either write them into the YouTube chat or you can ask them personally during the Q&A session. The link to the Q&A session will be posted into the YouTube chat just before it, so if you want to join the Zoom call, then mute your YouTube livestream and join with your full name. Now please let me introduce our today's speaker, Professor Martin Nieneck. Professor Martin Nieneck is a researcher at the Department of Biochemistry, University of Zurich. He completed his undergraduate studies in natural sciences at Trinity College, University of Cambridge, and he got his PhD at the European Molecular Biology Laboratory in Heidelberg. Professor Jinek conducted his postdoctoral research at the University of California in Berkeley with Professor Jennifer Doudna. They were studying CRISPR systems and their use as gene editing tools. For his research, the 2020 Nobel Prize in Chemistry was awarded to Professor Doudna and Professor Charpentier. In today's talk, Professor Jinek will be speaking about CRISPR-Cas genome editors. Please welcome Professor Jinek. Professor Jinek, the stage is yours.

Okay, thank you very much, Mirka and Indra, for the invitation to give a lecture within the framework of the Trinity College Science Society. I'm really happy to be able to do that. So yeah, hello and good evening from Switzerland. I will share my screen to— hopefully it's up now and it's working. So yes, thank you very much once again for the invitation. It was really a no-brainer to agree to give this talk, as I have a connection to Trinity. And so today what I will try to do is to give you an overview of what CRISPR-Cas genome editing is, and how we basically came to where we are now with the technology. How basically a discovery of the technology was, well, how the development of the technology was driven by discoveries initially made in bacteria and then in test tubes, and then led to the technological revolution that these new genome editing tools give us. So before I get on with that, however, I would again like to thank you for the invitation because I have a connection to Trinity. I was an undergraduate student at Trinity College between '98 and 2002. I must confess that there were no digital cameras and no smartphones in those days, so there are actually very few photos of me from that time, and the photo that you see on the screen is actually from a few years later when I came back to Trinity to pick up my free MA degree. This was back in 2004 or 2005, and I had a digital camera back then already. Okay, so anyway, so yes, it's an absolute pleasure to be back at Trinity, virtually at least, and to share with you my journey through CRISPR and through genome editing. As was already mentioned, I'm now based at the University of Zurich, and so what I would like to do here is basically acknowledge my research group. It's a wonderful group of PhD students and postdocs and also undergraduates who come and do their bachelor and master projects in our lab. And if there's one thing that I'm proud of in my research is it's really this, the fact that, you know, I have the privilege and the good fortune to work with really fantastic people. And so my research is basically firmly rooted in my long-standing passion for RNA, RNA as a molecule, and for its various molecular activities in cells. And in particular, I'm interested in the molecular mechanisms of protein-RNA complexes and various molecular machines that work on RNA. And this in particular in the context of genome defense mechanisms and their applications as genome editing technologies. And so where all of this basically comes together to a large extent is CRISPR, both as a genome editing technology, but also as a bacterial immune system that is used to find viruses in bacteria. And so as you're probably aware, we're really in the midst of a genome editing revolution, The emergence of the CRISPR-Cas genome editing technology is something that allows us to modify the genetic information in cells and organisms in a way that's simple, fast, and, you know, precise, and also quite, quite cheap, relatively speaking, is something that gives us unprecedented powers for not just basic research but also for biotechnology and for molecular medicine.. And in particular, the application of CRISPR-Cas genome editing for gene therapy is something that I think myself and a lot of people are of course very excited about. Another aspect of genome editing and CRISPR-Cas, and this is something that I'll hopefully come to in the last few minutes of my talk, is that You know, this technology is really giving us for the first time the power to modify our own DNA in a way that can be really transmitted to the next generation. And with this power, of course, I would say comes the responsibility to use this, to use this technology wisely. And so since 2012, when CRISPR-Cas genome editing emerged, this has really become a technology or a method for the masses, so to speak. Due to the reduced costs and the simplicity of the method, it has now basically thoroughly revolutionized not just basic research, but it's making a lot of breakthroughs in biotechnology as well as in molecular medicine, and really giving us the ability to do things in the genomes of cells and organisms that until a few years ago were not really possible. And so the, the basic principle of CRISPR-Cas genome editors is that at the core of the technology we have protein-based enzymes that act as DNA nucleases, as DNA-cutting enzymes, but whose specificity in terms of the DNA sequence that they recognize, that they can cut, is determined by the sequence of a guide RNA that can be made and then associated with the Cas nuclease. To program it to target a specific sequence in the genome of the cell. And so what basically these genome editors allow us to do is to cut DNA with quite a high degree of precision in cells or in organisms and generate something known as a double-stranded DNA break. And this is basically a type of DNA damage that the cells encounter all the time and have for that purpose evolved DNA repair mechanisms. And it is basically this combination of the DNA cutting and then the subsequent DNA repair that can be exploited for the, for the installation of genetic modifications in DNA, because the DNA repair mechanisms can be basically exploited and manipulated to some degree. To achieve specific DNA repair outcomes that result in modifications. So we have the combination of DNA cutting with endogenous DNA repair, which then leads to the modification that we would like to achieve. And depending on which cells we, we do this in, and depending on which DNA repair pathway can be exploited. We can basically do on one hand rather simple things like, you know, knocking out genes or making mutations at specific positions in the DNA, or do more complicated genetic modifications such as inserting DNA sequences into a specific locus in the genome. Or things like, you know, correct various mutations at a specific site in the genome. And so again, very, very simply put, the CRISPR-Cas genome editing technology really allows us to, you know, modify DNA as if we were basically modifying a piece of text on a computer. And so the discovery of the mechanism underlying these genome editor nucleases and their RNA-based programmability is something that, as was already mentioned, was recognized with the award of the 2020 Nobel Prize in Chemistry to Emmanuelle Charpentier and to Jennifer Doudna. You know, I was very privileged, I guess, and fortunate to be basically part of the team that was behind the work that led to the Nobel Prize. So in 2012, we were basically the team that characterized the first RNA-guided DNA-cutting enzyme and basically set off the genome editing revolution. On its trajectory. So what I would like to do today in this lecture is to basically give you an overview of CRISPR-Cas genome editing, about its mechanisms, about its applications. And so I'll begin by telling you a little bit about where CRISPR-Cas came from. I will then follow up by explaining a little bit about how these CRISPR-Cas genome editors, how these DNA cutting enzymes work at the molecular level, I will talk a little bit about how CRISPR-Cas genome editing can be implemented in cells and organisms, and then finally how it can be used to treat human diseases. And so I'll begin by telling you where CRISPR actually came from. And so as I already alluded in, at the beginning of my talk, CRISPR has its origin in prokaryotic cells, in bacteria and archaea, where they— where the system basically provides a kind of a genome defense mechanism, if you will, or an adaptive immune system that provides resistance against viruses, against mobile genetic elements such as plasmids and others. And the first sign, or the first indication, first, I guess, discovery of the CRISPR system was made back in the 1980s when the first genomic sequences of bacterial genomes became available, where it became clear that in bacterial genomes we basically have these specialized repetitive sequences consisting on one hand of basically repeats that had a specific sequence and a specific length, which were interspaced with non-repetitive sequences that became known as spacers. And so what basically defines the CRISPR system in a bacterium or in an archaea is basically this presence of this repeat spacer pattern in the genome. And this is basically then where CRISPR actually got its name, because it's based on this— it's an acronym based on Clustered Regularly Interspaced Short Palindromic Repeats, which basically describes the kind of the appearance of these loci in the genome. And so for about 20 years, the function of these CRISPR loci was really a mystery until in 2007 a group of researchers at a food company called Danisco made the discovery that these CRISPR systems actually functioned as an acquired resistance mechanism against viruses. So as a kind of a primitive adaptive immune system, if you will. And so What became basically clear was that these spacers, which I'm showing here as these colored squares in the bacterial chromosome, basically had sequences that were matching the DNA sequences found in viruses. And that the presence of a spacer in the genome of a bacterium was basically sufficient to make that bacterium resistant against the virus that contained the same sequence in its DNA. What also became clear around that time was that the CRISPR system was adaptive, meaning that bacteria would be able to take DNA from a previously unknown virus incorporate it into its chromosome, into its own genome, and acquire resistance against that virus. And so this mechanism basically can be essentially thought of as a kind of a vaccination card for the bacterium. Bacterium basically makes notes or makes a permanent record of its previous encounter with a virus in its own DNA, and then uses it to recognize the virus and to target it and eliminate it the next time that the bacterium is basically infected by that virus or comes into contact with it, with its DNA. Now, at the molecular level, this is where basically things got very interesting for me, because it became clear that there's RNA involved in the mechanism. So what happens is that this spacer repeat structure in the genome is transcribed as a long RNA precursor, and then processed into individual small RNA molecules, each of which basically then contains a sequence that's derived from one of the spacers.. And what these RNAs then do is that they associate with proteins that are produced by the transcription and translation of the so-called CRISPR-associated genes, which are always found together with this, with this array in a CRISPR system. And they form basically targeting complexes in which the the CRISPR RNA functions as a molecular guide. Basically, the RNA uses the information encoded by the sequence of the spacer to basically target the Cas protein to the DNA of the virus using base pairing interactions. And so in this way, the RNA basically provides the targeting capability and marks the DNA of the virus for destruction, typically by some kind of a DNA nuclease degradation mechanism. And so again, you know, if we want to kind of summarize the function of the CRISPR system, it's basically that first the bacterium will take the DNA of the virus and insert it into its own genome to, you know, effectively memorize the encounter with the virus. And then based on that information, it will generate guide RNAs that will program these Cas proteins for the recognition of the viral DNA by base pairing interactions. And this basically recognition mechanism is then coupled to DNA targeting by nucleases to basically destroy the DNA by cutting effectively. And so my contribution basically to this in the early days of CRISPR genome editing was that we, and by we I mean myself, Krzysztof Kelinski, and Jennifer Downer and Emmanuelle Charpentier and our colleagues, we discovered that in a subset of these bacterial CRISPR systems, we really had a protein known as Cas9 that combined basically all of these functions, in that it acted as a DNA-cutting enzyme that was programmable by these guide RNA molecules.. And that at least in its natural form, basically associated with two RNA molecules, one of them being the CRISPR RNA that provides the targeting information, and then the other being an RNA molecule known as tracer RNA, which basically acts as a kind of a coactivator. It's basically required for the CRISPR RNA to be recognized and to associate with the Cas9 protein and to program it. And so it's the combination basically of the CRISPR RNA and the tracer RNA and the Cas9 protein that then enables the bacterium to target the viral DNA that's complementary to the targeting sequence in the guide RNA. Now we realized in the course of our biochemical studies that we could basically take the dual guide RNA structure and we could simplify it by fusing the CRISPR RNA and the tracer RNA together to generate basically a single molecule guide RNA structure. And so basically reduce the complexity of the system even further to the point where we have basically just one protein to do the DNA cutting, and one guide RNA format with which we can program this protein to target the sequence that we want to modify. And so the idea of this one guide RNA, one protein system is then the basis of the genome editing technology as we know it nowadays. And so what I'll talk about now is basically how these genome editor nucleases actually function at the molecular level. And this is something that I personally am very interested in, and it's something that my own research group in Zurich has focused on over the last 8 years, where we were basically trying to understand the molecular mechanism of these nucleases using a combination of biochemical and structural approaches. And so we started working on Cas9 in this respect, because this is, at least mechanistically, this is a fascinating enzyme. So as already mentioned, it basically has this dual RNA guide structure. That it uses to find the matching sequence in the target DNA. And this is basically dependent on the base pairing interactions between a 20-nucleotide sequence in the guide RNA. So 20 letters of the RNA have to match 20 letters in one of the strands of the DNA.. And if this is the case, then the enzyme will basically catalyse the cleavage of the DNA at this position within the target site, and it will do it by basically independently catalysing the hydrolysis of the two DNA strands. And for that, it has two, two nuclease domains, so two active sites in this enzyme to catalyse the the two DNA strand cleavage events. Now this is actually a little bit more complicated because the process of DNA targeting and cutting also requires that the DNA contains an additional sequence motif outside of the target site, and this is referred to as a something called protospacer adjacent motif, or PAM. And for the most widely used Cas9 genome editor, this is basically a sequence of this consensus, so NGG, where N can be any of the 4 DNA bases. And so this has to be present basically in the DNA, otherwise the system will not recognize the target and will not be able to cleave it even if it's perfectly complementary to the, to the guide RNA molecule. So this is an additional requirement basically for the DNA cutting. And of course, to be able to recognize the DNA in the first place, you have to have the match between the guide RNA and the target DNA, particularly in this, you know, PAM proximal side of the duplex. But this in itself is not, again, not enough. It's sufficient for DNA binding, but if you need to cleave the DNA, then you also need to have base pairing in this part of the RNA-DNA duplex at the other end. And again, there's a there's an explanation for why this is in the molecular mechanism of the enzyme. And so we were very interested in these aspects of the targeting mechanism, and we wanted to gain insights into this. And in my lab, we do this by basically taking a structural approach. We use structural methods like X-ray crystallography, and nowadays also increasingly cryo-electron microscopy, to basically determine structures of these enzymes in various states, and then to basically, you know, learn more about, about the molecular mechanisms that we can then test using, using other approaches, either biochemical assays or biophysical assays. And so on. And so our work on Cas9 basically led to one of the first structures of the protein in complex with the guide RNA and the target DNA. And what I'm showing here is basically a 3D printed model of the molecule. In white we have the Cas9 protein, in red we have the guide RNA, and then in the two shades of blue we have the two strands of the DNA target. And so what you see here in the middle is basically the RNA-DNA duplex that the enzyme makes. And that's positioned right in the middle of the molecule, kind of surrounded by various parts of the Cas9 protein. One other interesting thing that you maybe can see here is that when Cas9 binds to DNA, it actually bends it by almost 90 degrees. And that's also something that we think contributes to this process of separating the two strands of the DNA from each other and forming the RNA-DNA duplex. And so the beauty of having molecular structures— and by the way, this structure was determined by my very first PhD student, Caroline Anders— is that you can really zoom in and get all the details. And so in our case, one of the details and one of the highlights of the structure was basically the mechanism by which this enzyme recognizes this PAM, this critical element of the recognition mechanism, where we could basically show that this was done by specific amino acid residues in the protein contacting the DNA bases using hydrogen bonding interactions. And so we've continued studying Cas9, But we also, in addition, started working a few years later on Cas12a, which by that point emerged as an alternative genome editor nuclease. Again, it's an RNA-guided DNA-cutting enzyme that is found in a different class of CRISPR-Cas systems.. And it has somewhat complementary and, again, very interesting properties. So unlike Cas9, it has a slightly different PAM specificity. The PAM is located at the other end of the target site with respect to the polarity of the guide RNA. It is a naturally occurring single RNA-guided system, so there's no tracer RNA in the Type V CRISPR-Cas systems where Cas12a is found. It cleaves the DNA by a slightly different mechanism, because it only has one nuclease active site. And so again, there are very interesting aspects of the mechanism that we then continued studying using structural and biochemical approaches.. And so one thing that I wanted to share with you is again, in this case, a movie that's based on a couple of X-ray crystal structures that we solved in my laboratory. This was the work of a postdoctoral fellow called Dan Swartz, and based on which we were able to again make a kind of a somewhat idealized movie of depicting how the DNA targeting mechanism basically works. And this is hopefully something that you can see. So we have basically the Cas9 protein and the guide RNA. And now the enzyme is basically depicted as basically sliding along a double-stranded DNA to search for its target. And the search basically begins with looking for the PAM sequence, and once the PAM sequence has been located by the enzyme, this basically triggers the unwinding of the two DNA strands and the formation of the RNA-DNA duplex, which basically forms in this kind of a zip-like manner from one end to the other. And when this occurs, then the DNA is basically then positioned within the protein in a way that the protein can catalyse the cleavage of the two DNA strands in an ordered sequential mechanism. And so by basically elucidating the structures of these enzymes, we have been able to, you know, highlight some of the mechanistic differences as well as commonalities that these two molecular systems have. And our goal in this whole enterprise is to basically be able to use these insights not only to improve our understanding of these systems in the context of the biology of CRISPR-Cas systems in prokaryotic cells, but also as a way to contribute to the continued development of CRISPR-Cas genome editing technologies. We hope to be able to use structure-guided engineering of the proteins as well as the guide RNAs to address some of the constraints that these systems have. Things like improving the specificity of these systems and minimizing their off-target activities, expanding the targeting space by basically modifying the PAM specificities of these enzymes, which, you know, constrain these enzymes to only specific sequences in the genome, and to do other things like, you know, improve the efficacy of these enzymes in specific cell types and and others. And so how do we actually go on about implementing CRISPR-Cas genome editing? And so again, I will come back to the, to the slide that I showed you at the beginning where I told you that what we basically do in genome editing is to exploit the RNA programmability of these DNA nucleases to make double-strand breaks, and then relying on the endogenous DNA repair mechanisms to then introduce modifications into the DNA. And again, very, you know, sort of somewhat simplified way, we have two main DNA repair mechanisms, one called non-homologous end joining, which can be used to make either short insertions or deletions in the DNA, because this is a kind of a quick and dirty mechanism that doesn't really care about whether the double strand break is repaired with any degree of precision or not. And so this can be basically used to make indels in the DNA, or to knock out genes by basically disrupting the open reading frame on the protein coding gene. On the other hand, we have homology-directed repair, which is a set of DNA repair mechanisms that rely on having a homology template for the repair process. And so by providing that repair template exogenously together with the Cas9 protein and the guide RNA, you can basically trick the DNA repair mechanism to use that DNA template, and to basically modify the DNA if that template is designed in a way to, you know, insert a piece of DNA into this, into the genome, or to overwrite part of the genetic sequence with something else. And so by combining basically Cas9 or Cas12a-based DNA cutting with homology-directed repair, we can, we can really do quite complicated and very precise genetic modifications. And so this idea of using DNA-cutting enzymes to do genome editing is actually something that was already known before CRISPR came along, and there were already quite powerful technologies to do that, but their main drawback was that they were based on designed, artificially engineered proteins in which a DNA nuclease, a non-specific DNA nuclease, was fused to sequence-specific DNA binding proteins, or to combinations of sequence-specific DNA binding protein domains. And so the main drawback of these two technologies, zinc finger nucleases and TALENs, was that, you know, every time you wanted to target a different sequence in the genome, you had to design a new protein. And even though the design of these proteins was quite modular, this was still a tricky thing to do. And so when CRISPR-Cas9 and and later some of the other genome editors came along, this process was greatly simplified because all you had to do now was to simply change the sequence of the guide RNA to reprogram the enzyme to target a different sequence in the genome. And so very shortly after our discovery in 2012, a number of studies were published showing that this could indeed be done in eukaryotic cells. And we were also involved, or I was also involved, in one of these studies showing that indeed Cas9 could be used together with these guide RNAs in eukaryotic cells for genome editing. So in our study, what we did was we basically took pieces of DNA, so DNA plasmids encoding on one hand the Cas9 protein, and on the other the guide RNA, and introduced them into cultured human cells in a, in a Petri dish. And a few days later, we basically isolated the genomic DNA from these cells and started sequencing individual clones from this DNA. And we were basically starting to see these telltale signs of DNA repair happening exactly at the sites that we were targeting with the enzyme. So we were basically recovering clones where part of the DNA sequence was deleted, as well as we had one clone where there was an insertion at this site. And so this was basically for us the first indication that we were getting DNA repair following DNA cutting by the RNA-guided Cas9 protein in cells. And so this then basically led to the implementation of the technology, so that basically nowadays all you need to do really to target a specific gene in the genome is to essentially modify the 20 nucleotides at the 5' end of the guide RNA, just basically writing the sequence that you want the genome editor to recognize in the genome and deliver the protein and the guide RNA, potentially also together with homology repair template into cells. And for this you nowadays have actually various options. The simplest way, as I already told you, was to simply deliver both components in the form of DNA that's encoding them and then relying on the cells to basically express the protein and the guide RNA to form the complex inside. For certain organisms or certain experimental systems, you can also deliver these components in the form of RNA. But something that's becoming more and more common in the field is basically to simply pre-assemble the protein-RNA complex in, in a test tube using a recombinant Cas9 protein that you can buy from a company. And a synthetic RNA that you can either make yourself in a laboratory or just buy again from a company that's specialized in oligonucleotide synthesis, and then delivering these pre-assembled complexes into cells using various methods. If the cells are big enough, you can do this by microinjection, but typically for cultured cells you would nowadays do it by electroporation. Delivering an electric pulse to the cells to temporarily disrupt the membrane and make them, make them take up the protein-RNA complex. And so this direct delivery of the protein-RNA complex is something that is quite nice because first of all you're not delivering any DNA into cells, So both the Cas9 protein and the guide RNA basically have a limited lifespan in those cells, and that means that you also end up with fewer off-target effects, because there's basically less time during which the Cas9 protein and the guide RNA are active, and during which they can not only modify the target site that you want to actually modify, but also potentially or accidentally modify the, you know, other sites in the genome that are partially complementary to the guide RNA. And so the RNP delivery not only is useful for reducing this off-target activity but also can improve the efficiency of the process to begin with, because the protein-RNA complex is active from the moment it gets into cells. And so nowadays, you know, if you want to basically modify the genome of a cultured cell line, what you need to do is basically get your recombinant Cas9 protein, design a guide RNA based on the sequence that you want to target, so you would look for the presence of a PAM, You would basically then take the sequence of the DNA next to the PAM, plug that sequence into the guide RNA, have that guide RNA made, and then deliver it into cells together with the Cas9 protein, as well as with any kind of DNA repair templates that you want to use for a specific modification, and then treat the cells using this electroporation approach. So put cells into this cuvette and then simply zap it with electricity, again very simply put. And if you do this and you design the repair template that in a way that, you know, not only introduces a piece of sequence into the genome but also introduces a restriction endonuclease site, you can then very simply you know, check for the presence of that piece of DNA and for its insertion in the DNA by isolating the genomic DNA of the cells and just doing a simple restriction digest. And so, as again, as I already mentioned, you know, when we think about implementing CRISPR-Cas genome editing there are still some constraints that we would ideally like to be able to address and to improve the basic capabilities of the technology. And in particular, at least in, you know, in the context of my work, we would really like to provide structural insights that can be used to do this. And so one of the concerns surrounding the use of Cas9, particularly in the context of genetic therapies, is the specificity. Already quite, quite early on, it became clear that these genome editor enzymes are perhaps not as specific, not 100% perfect, and that they basically can target other sequences that are similar but not perfectly matched to the guide RNA. For Cas9, sequences containing up to 5 mismatches can be tolerated by the enzyme to some extent. And so there's quite a lot of effort now in the field aimed at improving the specificity of the enzyme and to basically reduce its off-target activity. And this is something that has led to the development of, you know, engineered high-fidelity variants of the basic naturally occurring enzyme, where by targeting specific amino acid residues in the protein that would normally contact the DNA, you can basically make the system much more sensitive to mismatches between the RNA and between the guide RNA and the DNA. By basically perturbing the, let's say, the thermodynamics and the kinetics of the DNA cutting reaction. And so by now there are basically several of these high-fidelity enzymes that are available and that have reduced off-target activity. Another development in the field is basically the emergence of CRISPR-based molecular tools that enable you to do genome editing without having to cut the DNA in the first place. And this is, this is a technology that was developed in David Liu's lab primarily at Harvard, where the technology basically relies on the fusion of Cas9 protein with an enzyme that can basically catalyze the chemical conversion of one base into another. So in this case, the Cas9 protein is not used to cut the DNA, but is used in its inactive form just simply as an RNA-guided DNA binding platform that will deliver this base-modifying enzyme to the DNA. So what this cytidine deaminase does, it would, it will basically take a C and turn it into a U, such that then when the DNA is replicated, that U will be decoded and, you know, and then A will be inserted in the opposite strand during DNA replication. So we can basically use this to turn one base pair into another in the DNA. Without the need for DNA cutting. Another recently developed technology again attempts to do editing without cutting by fusing again a nuclease-deficient form of the Cas9 protein with a reverse transcriptase, and using a modified guide RNA as a template for the reverse transcription reaction where the DNA, once it's been nicked by the Cas9 protein, will be extended by the reverse transcriptase activity and then lead to, you know, insertion of sequences into the site based on the sequence of the modified guide RNA. So this approach has been it's been termed prime editing, and again it's a, it's again another development in this molecular toolkit for genome editing. And so finally, what I would like to talk about a little bit is about the potential of CRISPR-Cas for genetic therapies.. And ultimately the idea here is that we could use CRISPR-Cas genome editing tools to really correct disease-causing mutations in the human genome, and to do this, you know, very precisely and without having to rely on, you know, kind of the previous approaches in gene therapy which relied on gene augmentation, basically on this idea that you know, we would essentially rescue the function of a defective gene by inserting another copy of the gene somewhere in the genome. So instead of this, what CRISPR-Cas would potentially allow us to do is to simply go back to the defective genome exactly at the site— sorry, go back to the defective gene exactly at the site in the genome where it's found, and just simply correct the disease-causing mutation. An example, a classic example of a situation like this is, you know, the sickle cell disease where a single nucleotide mutation in the beta globin gene leads to the production of a defective hemoglobin variant that, you know, causes the disease phenotype. And so one could basically approach this by simply correcting that mutation in that gene. And so for this, again, one can think about two approaches to do this. One would involve delivering the CRISPR-Cas9 or CRISPR-Cas12a or any of these technologies directly into cells or organs or tissues in a living human patient. So basically doing in vivo genome editing, and this is something that is being developed and where the technologies basically will use things like viral vectors or lipid nanoparticles to basically deliver these components into cells. Another approach which is proving to be a bit more simple, and at least in the short term, is ex vivo genome editing, where the goal is to basically do genome editing in cells that have been extracted from the patient, modified in a laboratory, and then reintroduced into the patient. Once the genetic defect has been corrected in these cells. And here in the future, what is going to be potentially very powerful is the combination of genome editing and the iPSC technology, so induced pluripotent stem cells, where, you know, we could conceivably, you know, take fibroblasts or other cells from the patient reprogram them to make iPSCs in the laboratory, then modify those iPSCs, differentiate them into the specific cell type that you would like to, like to, like to get, and then introduce those cells back into the patient. So there's a lot of, of course, effort now focused on implementing these two approaches for specific genetic diseases. So I will give you a few examples of this. So there are by now, you know, dozens of clinical trials using CRISPR-Cas genome editing. This is an example very recently of a genome editing strategy to treat beta, beta globin deficiencies. Such as sickle cell disease or beta thalassemia by basically disrupting the beta globin gene which carries the mutation and basically reactivating the expression of gamma globin, so the fetal globin subunit, in adult erythroblasts. And so this is something that is already in clinical trials and It will be of course very, very interesting to see how this will develop, and I think the, you know, the initial results are incredibly, incredibly promising. Another area where ex vivo genome editing in cells is making a big impact is in the process of engineering immune cells for treating various forms of cancer. So in this case, the goal is not to correct, you know, a genetic defect, but rather to reprogram the immune cells of a patient to basically make them recognize cancer cells specifically. So this is basically by inserting specific receptors on the surface of these, of these T cells, and also by, you know, introducing other genetic modifications that make these T cells, you know, irresponsive to various forms of inhibition and so on. And so reprogramming T cells to target cancer cells is going to be again another powerful technology and powerful application of CRISPR-Cas. If we talk about in vivo editing, So here again, there's quite a lot going on. I think overall the field is maybe not as advanced as for ex vivo genome editing on the whole. But there have been a number of really breakthrough studies where by delivering CRISPR-Cas components into model organisms, which serve as, you know, models of human genetic diseases, you know, other researchers were able to really get some spectacular results. An example of this was basically these 3 studies showing that in a mouse model of Duchenne muscular dystrophy, you know, a really debilitating neuromuscular genetic disease, you know, the treatment basically with with virally delivered Cas9 was able to basically restore muscle function in these animals. And so perhaps unsurprisingly, you know, there's a lot of excitement about the prospect of using genome editing for therapeutics. There are a number of companies that are active in this area and, you know, a number of different diseases that are being actively investigated. And so it's going to be something that will be, you know, coming to the hospital near you very soon. And this is also reflected, I guess, in the enthusiasm of the of the stock market for these companies. But before we can get there, I think we will have to address some of the constraints of the technology and of its specific applications. So the ongoing work in the field is really focused on things like improving the specificity on improving the delivery. For many of the diseases and many of the indications, the delivery will actually be the more tricky part of the therapeutic strategy than the actual editing. But also for the editing, one of the key areas is to really get a better way of controlling the DNA repair process, or to avoid the endogenous DNA repair process altogether by using some of these technologies that don't necessarily rely on them in the first place. And so this is something that, you know, is going to revolutionize genetic therapies and molecular medicine in the future. And so finally, something that I want to mention at the very end is, is the ethical side of it, because until now we've talked about, you know, editing human patients and basically doing the editing in somatic cells. But of course there's, there's the possibility of using CRISPR-Cas genome editing also for editing in the human germline, and this is something that's, you know, a very ethically controversial aspect of CRISPR-Cas genome editing. This of course became the topic quite early on in 2015 when the first genome editing experiments were done in human embryos, and then again was reinforced by the reports in 2018 of the first genomeditated babies being made in China. And so of course this is something that is creating a lot of controversy in the field, and it's, it's what I, what I see as a, as a good development is that there is basically an ongoing debate about what would be the responsible path forward for the use of genome editing. There have been, you know, global international summits and ongoing discussions. It's clear that for now, you know, research and the use of CRISPR-Cas genome editing for somatic gene therapies is something that should go ahead. The use of germline genome editing is something that should not be done. For two main reasons. One, that there is currently— well, the technology is not mature enough yet to ensure that it can be safely and efficiently used, but mainly because so far there hasn't really been a widespread societal consensus on the use of this technology for germline genome editing. In many parts of the world, this is something that's prohibited by law, and, you know, not possible. And so it's clear that we will have to have ongoing discussions, and it's something that, you know, the scientists cannot really decide for themselves. This is something that really involves the whole society and all the different stakeholders in this. So with this, I would like to conclude my talk. I hope that I've given you an illustrative overview of the journey that CRISPR-Cas has taken from bacteria all the way to biotech and to micro-medicine. And I would like to basically finish on this note by saying that, you know, the discoveries that led to the development of genome editing, this was really basic research. This was not something that initially aimed to do genome editing. This came out of, you know, people being curious about bacteria, about how they defended themselves from viruses, and how these, you know, how these sophisticated molecular systems actually functioned. And so if there's basically a lesson in this, it's this: that if you think you have a great idea for a technology, nature has probably been there before you. And so it's, it's worth basically investing into basic science to, you know, to look for interesting and potentially useful molecules in nature to then repurpose for other, for other applications in technology. So with that, thank you, thank you very much for listening. I'll be happy to answer any questions and Yeah, once again, thank you to the organizers for inviting me to give the talk at Trinity College. It's great to be back after, after all these years.

So thank you. Professor Iñak, many thanks for the great talk. And now it's time for a Q&A session. So Indra, can you please post the link? Yeah, the link to join our Zoom call is in the YouTube chat, so please mute the YouTube live streaming beforehand and join with your full name if you have any questions.

And meanwhile, there are some questions. Okay. Um, hello, can I fire off? Or sure. Yes, hi. Yeah, thanks very much for your talk, Martin. I was wondering about the mechanism of Cas12 binding to DNA, because you showed that for Cas9 there are some arginine fingers binding to the PAM motif. So is it conserved?

Is it the same in Cas12, this mechanism? The mechanisms are perhaps unsurprisingly different, given that the PAMs are also different. In Cas12, the mechanism is some— is actually based more on the readout of the shape of the DNA rather than on the actual recognition of specific bases in the PAM. So for Cas9, the PAM is T-rich, and, you know, if you have a stretch of, or a series of, T residues, or AT base pairs, then the minor groove of the DNA is basically a bit more narrow. That's something that the enzyme can recognize, because then it has a shape that's basically complementary to the shape of the DNA. And this is basically why Cas12 enzymes in general actually prefer these. In other Cas9 proteins, that have different PAM sequences. There you again see different amino acid residues contacting the PAMs and different combinations. And this is something that's also been exploited to actually engineer new variants of Cas9 that recognize other PAM sequences.

So you are looking at the different evolutionary solutions in the lab as well, like pursuing the catalytic mechanism and trying to improve the current state of the art?

So we already looked at the PAM recognition mechanism earlier. So back in 2015, the first engineered variants with different PAM specificities were developed. And then very shortly after that, we determined structures of those. And that structural information kind of fed the development of yet another generation of genome editing variants of the kind, not in my laboratory, but in other laboratories. And so our current focus or current interest is actually the off-target activities. We're actually very interested in, you know, what it means when you have a DNA that's not perfectly copied. Complementary, and how come, you know, how come this enzyme actually is quite tolerant of mismatches in certain parts of the RNA-DNA duplex, and what kinds of mismatches, and so on.

So this is something that is an ongoing project in the lab. Well, on that note, I was going to ask about your opinion on the prime editing technique, because I was trying it in the lab during the summer, and I got like almost zero efficiency in editing with this. And it also looks a bit dodgy because there are so many improbable events that need to occur before you actually get the edits.

So do you think this is usable in like everyday lab work or stuff like that? So, you know, I don't have a direct experience with this method. I'm more of a basic researcher than an audio developer, but the feedback from my interactions with some of my colleagues who have used this, it seems to work quite well in certain cases and sometimes it doesn't. And right now it's still too early to tell why exactly. I see, yeah. I've heard, you know, people having quite good experience. Maybe it's a bit, you know, a bit dependent on the specific target and, you know, specific cells and so on.

Okay, thanks. Thanks very much. And then perhaps final question. I'm trying— I find it difficult to conceive of how CRISPR binds in the context of chromatin. Because if you think about it in bacteria, it's obvious that there is no— there are no nucleosomes and stuff like that. So it can bind directly to viral DNA invading the cell. Whereas in mammalian systems, for instance, you've got the repressive effects and the environment of chromatin. So how does it overcome it? Or if you know the kinetics of how it actually scans the DNA, I haven't like perused the literature on that.

So this is something that has not really been completely resolved yet. At least in vitro, it's clear that nucleosomes actually are quite refractory to Cas9 binding to the DNA and cutting it. Actually, if you want to cut DNA that's within chromatin, at least in vitro, you have to make the enzyme, you know, target essentially the linker DNA between individual nucleosomes. I think what happens in cells is that, you know, the chromatin is actually not static. It's quite dynamic, it's remodeled. Of course, in some parts of the genome, you know, more than in others. But at some point, you know, the DNA site that you want to target basically becomes exposed and essentially nucleosome-free, if you will. You know, especially if you think about proliferating cells where you have to replicate DNA anyway, and you know, chromatin is remodeled in the course of DNA replication and so on. So it's something that, you know, is probably a factor here. Because at least in vitro, if you have nucleosomes on the DNA and they're kind of bound at a specific place, then it's actually quite hard for Cas9 to access the DNA. So do you think if you fuse Cas9 to some nucleosome remodeling active-like enzyme, could you improve the efficiency Well, this is something that people have been trying to do, yes. Okay. Yeah. Also as a way of controlling transcription of genes, basically, to transcriptionally activate or repress a locus by delivering either chromatin remodelers or nucleosome-modifying enzymes.

Yeah. Basically, methyltransferases, acetyltransferases.

Okay, thanks very much for your answers. Cheers. Fritz, now we have a few more questions in the YouTube chat. So first one is: how do you make the cell do homologous recombination to repair the DNA when you want it to repair it that way? Or how do you make the cell do NHEJ to repair the DNA when you want to do gene knockouts?

Right, so this is something that you don't necessarily have a lot of choice about, at least, you know, without anything, you know, without the help from something else. In most cells, the non-homologous unjoining is basically the predominant DNA repair pathway. The homology-directed repair is, at least in my understanding, is something that's really only active in in proliferating cells where DNA is being replicated. Because, you know, the purpose of— you know, you can only really repair by homology if you have the template to begin with, and, you know, you have the template basically when you've replicated your DNA. So homology-directed repair is of course most active in proliferating cells and is less active or completely absent from post-mitotic, you know, differentiated cells. So this is why it's, for example, very difficult to do homology-based editing in things like neurons or, you know, other differentiated cell types. And so you can improve the efficiency of homology-directed repair using several tricks. One is, for example, to use chemical inhibitors or, you know, other ways of inhibiting or repressing non-homologous end joining. Another way is to, for example, synchronize the cells and basically deliver the Cas9 guide RNA at a specific point in the cell cycle. There are other ways, you know, we had a collaboration with the lab of Gerald Schrank here in Zurich where, you know, we attempted to improve the process by basically covalently linking the Cas9 protein and the homology repair template, and this also has an effect. So there are ways to bias this towards, you know, one repair pathway or the other, but still this is something that's a bit of an unresolved problem with the technology, or one of the limitations.

Great. Then there was a subsequent question from the same person also. What specific function does the PMA sequence actually have in relation with the Cas9? I understand that PMA is required to cleave DNA, but why?

Well, it's basically required because when the enzyme binds the DNA, it looks for the PAM first. And the PAM really establishes the, if you will, the register of the DNA sequence against the guide RNA. Basically, when Cas9 or Cas12 have located the PAM, then We still don't know what happens in the, really in the very early stages of this process, but the assumption is that this initial interaction with Pam either distorts the DNA or does something to the DNA so as to facilitate the initial separation of the two strands, so that once the strands have been separated, the target strand can start base pairing with the guide RNA. Because I mean, the problem is that, or, you know, the beauty of these enzymes is that they essentially unwind the DNA without any energy input. You know, they just use the thermal energy in the system and then rely on the fact that, you know, overall the process is kind of thermodynamically downhill, that for every base pair in the DNA that you break, you make a base pair with the DNA. And so if you do this, you know, one base pair at a time, you know, the overall, say, activation energy barrier is lowered. But still you have to get to the first opening of the DNA, and this is basically coupled to the process of PAM recognition. So the PAM is basically necessary to open up the DNA and then to form the structure that allows the the guide RNA to base pair with the DNA.

Great, then another question: What is the main principle behind the off-site targeting? Is it the non-perfect homology between the guide RNA and the target sequence, or the Cas enzyme itself?

So this is to a large extent basically driven by the tolerance for mismatches between the RNA and the DNA at specific positions. And also the type of the mismatch seems to have an influence on how bad, you know, how bad the off-target activity is.

Then another one, I think it's maybe related to the—

does DNA modification or accessory proteins affect the targeting of the gene? Well, it's a bit of a yes and no question. So at least the, you know, the DNA modifications that we encounter in eukaryotic DNA, like, you know, cytosine methylation and so on, this doesn't really affect the basic recognition mechanism. There are, however, certain DNA modifications that viruses can make to their own DNA that may— that can make the DNA refractory. So there are certain bacteriophages, they glucosylate DNA bases, and depending on the position of the glucosylation, you can inhibit Cas9 activity. As it turns out, Cas12a is not so sensitive to this, and there is, we think, a sort of a structural explanation for it, because, you know, it all comes down to how much room there is in the major groove of the DNA, and whether that, you know, somehow sterically clashes with the protein and so and so on. But for the eukaryotic modifications, this is not a major factor, as far as I'm aware.

And last question from the chat: could you comment something about non-biomedical applications of CRISPR-Cas?

So this is something that I obviously haven't really had the time to go into, but that's Again, you know, another big area of interest where, you know, one could think of using Cas9 in, you know, synthetic biology to really, you know, generate new microbes that, or genetically engineered microbes that would have interesting properties, you know, be able to, you know, do biosynthesis of new molecules and so on. So that's something that genetic engineering— I mean, in many organisms, the genetic engineering is already possible now. But there are organisms where this could make a big difference and is being used to engineer biosynthetic pathways and so on. So that's one area. And the other is, of course, modifying potentially plants or animals for use in agriculture. And again, there's quite a lot of interest, quite a lot of development. But again, you know, this is also something where there's quite a lot of controversy, particularly in Europe, and, you know, in the context of, you know, would CRISPR You know, would organisms modified by CRISPR, for example, fall under, you know, the current GMO regulations, and so on. And where I think there's a bit of a difference between Europe and the US and the rest of the world in how this is being approached.

So Lukasz, do you have many, any more questions, or? Um, yeah, can I have one more question please? So you showed us the crystal structure of Cas9, and I mean, if I look at it, it seems like there could be many heterogeneities introduced by the differences in the DNA positioning and RNA positioning. So I was wondering how you made it crystallize in the first place, or if there were any issues with that, or if you— how did you stop it at the intermediate which you're like showing in the crystal structure?

Um, you know, so to make Cas9 crystallize, first of all, you know, we were trying to prevent the DNA from being cut, so we were using a catalytically inactive mutant, but you can also crystallize DNA with the catalytically active protein. The issue is that, you know, if you want to crystallize something, you have to— I mean, the molecules will crystallize only if they form contacts, you know, that lead to the formation of a crystalline lattice. And this is where, you know, this is a process that's not so easy to control. You can, you know, you can try to facilitate it by, especially if you work with nucleic acids, by, you know, engineering the DNA and the RNA in a way that, for example, you know, makes sticky ends where the bases could, you know, somehow come together, and this would lead to the formation of crystal contacts somehow. I think, you know, in, for the Cas9 structures, we were actually quite lucky that, you know, the specific sort of DNA design that we took just accidentally led to the formation of a crystal contact between one loop in the guide RNA and basically the part of the one side of the duplex where there's basically a blunt end, and you have bases and base pairs kind of stacking on top of each other. For the Cas12 structures, there we had to use a trick and really modify the DNA to have overhangs. And this led to the formation of basically a kind of a continuous stack of molecules along one of the directions of the one of the axes of the crystalline lattice. So this is something that, you know, you kind of had to do in X-ray crystallography, but nowadays, you know, with the advent of cryo-electron microscopy, this is something that we don't really have to worry about anymore. And so this is why cryo-EM has been so powerful also in the CRISPR field. Because these molecules are, they're big enough to be imaged in the electron microscope.

So this is where everybody's kind of moving, including us.

Okay, thank you very much. Looks like there are no more questions, so Professor Jeneck, many thanks for your great talk again and for answering the questions.

Hey, thanks so much. Once again, also for the invitation. It's been a great pleasure, and yeah, hope to have a chance to see you at some point. Yeah, we'll be looking forward to it when conditions permit. I haven't been to Cambridge in a few years, so it would be great to come back.

All right, take care. You too. Bye-bye. Bye-bye. And next Thursday there will be a talk by Professor Harry Anderson from the University of Oxford, and Professor Anderson will be talking about molecular nanostructures and their unusual properties. The talk is going to start at the usual time at quarter past 6 of the UK time. So thank you for attending the talk and have a nice evening.