Scientist Stories: Jennifer Doudna, Genome Engineering and CRISPR

Jennifer Doudna was born in 1964 in Washington, D.C. In 1989, she received her PhD from Harvard Medical School in Boston. She currently serves as a professor at University of California, Berkeley in the U.S. Jennifer Doudna, I now welcome you onto the stage. We are very much looking forward to hearing your lecture.

I'd like to begin by thanking the Royal Swedish Academy of Sciences, the Nobel Prize Committee in Chemistry, my family, including my spouse, Jamie Kate, our son, Andy, my sisters, Ellen and Sarah Doudna, my friends, colleagues, and of course, my former and current lab members about whose research I will be speaking today. It's a wonderful honor to have this opportunity to share with you the science that we've done over the last few years and to discuss the extraordinary opportunities and exciting advances that are happening right now with CRISPR-Cas9 as a genome editing technology. I thought I would begin by telling you about the origin of the ideas around CRISPR, and this began at least two decades ago with research in microbiology laboratories showing that bacteria might have an adaptive immune system, a way to provide protection against viral infection that would allow cells to acquire it in real time and then protect the cells from those viruses using this recording system in the genome. And this system came to be called CRISPR, and as I'm showing here, this is a mechanism by which bacteria and archaeal cells can adapt to viruses by integrating small pieces of viral DNA into the organism genome at a site called CRISPR, and then using that information to make an RNA copy of the sequence that can then provide the information required for detection and cleavage of viral DNA, or sometimes viral RNA, using CRISPR associated, or Cas proteins. And this slide shows the way this system is thought to act in bacteria and represents experiments that were done in the field early on, both using bioinformatics and molecular genetics to understand the function of this pathway. This video shows the way that we imagine CRISPR operates in nature, where bacteria that might be growing in a biofilm are being infected by viruses. And when the virus injects its DNA into the cell, the cell can acquire a small piece of that foreign DNA into the CRISPR locus, a place in the genome that consists of repeated sequences that flank these integrated sequences from viruses that are called spacers. And the cell then makes a copy of that integrated sequence in the form of RNA, which is then processed into individual units that include a sequence derived from a virus. Those RNAs combine with a second type of RNA called tracer and a protein called Cas9 to form a surveillance complex, an RNA-guided protein that can search the cell looking for sequences that match the sequence of the guide RNA. And when a match occurs, the CRISPR-Cas protein, Cas9, is able to cut double-stranded DNA, triggering destruction of the DNA in a bacterial cell. And this pathway has been operating for a long time, over evolutionary time, evolving in bacteria, and has, over time, been diversified into many different forms of CRISPR. Systems that can operate effectively in these organisms. But today I'm actually going to talk about, in particular, about one type of CRISPR-Cas system that uses the protein Cas9, as I showed in that example. This slide shows the three steps in CRISPR-acquired immunity that include adaptation, expression, and interference. And in the research that I did in my laboratory with a number of my former students, beginning with Blake Wiedenheft and then Rachel Horwitz, and folks that came after them, we began investigating, in particular, this third step in the pathway called interference, the stage of the CRISPR pathway that involves an RNA-guided detection of viral DNA. Back in 2011, we had the good fortune to begin a collaboration with Emmanuelle Charpentier and her student, Krzysztof Chylinski. And this launched a wonderful opportunity to answer what was at the time a very interesting and intriguing question in the CRISPR field, namely, what is the function of the protein called CRISPR-Cas9? And we were fascinated by this question because this particular protein had been implicated in protection of cells, and in particular, a type of bacteria that Emmanuelle's group was studying called Streptococcus pyogenes, an organism that infects human beings. And this particular organism has a CRISPR-Cas9 protein encoded in its CRISPR system that was implicated in protection of cells from viral infection. But the question was how. And so in our collaboration, we investigated this and addressed the question by doing biochemical experiments that involved working with purified CRISPR-Cas9 protein. And the RNA that guides it to target DNA sequences in cells. And what we found, and this is research done by Martin Jinek, a former postdoc in my laboratory, and Chris Chylinski working in Emmanuelle's laboratory, is that CRISPR-Cas9 in nature is a dual RNA-guided protein. As I showed in that introductory video, this is a protein that uses a CRISPR RNA molecule to direct it to a sequence of DNA that matches the CRISPR RNA sequence, and it also requires a second RNA molecule called tracer that provides the— an interaction with the CRISPR RNA required for assembly with Cas9. And so together, these two RNAs guide Cas9 to DNA sequences where the protein Cas9 is able to make a cut in the double helix of DNA. I think one of the wonderful aspects of CRISPR and this project with Emmanuelle's group is that this was the point in our research where what began as a curiosity-driven investigation morphed into a project that had much broader implications. Because once we understood how nature uses a dual RNA system to guide Cas9 to target DNA sequences, it was possible to engineer the dual RNA guide as a single guide RNA, shown here, that included both the targeting information and the structural requirement for assembly with the Cas9 protein in a single RNA molecule. And once this experiment was done by Martin and by Christoph, we found that Cas9 could be programmed with single guide RNAs and directed to cleave double-stranded DNA at a desired sequence, taking advantage both of the actual targeting information in the CRISPR RNA molecule, as well as the requirement for what's called a PAM sequence, a protospacer adjacent motif. PAM is easier. That occurs right next to the target sequence in the DNA. And those two elements are required for Cas9 to recognize and cleave DNA. But this requirement is sufficient for Cas9 to function as an RNA-guided DNA cleaver. And I wanted to show you a key experiment that Martin Jinek did to test this idea of a single-guided Cas9 protein that could cleave different sequences of DNA. And the experiment was to design guide RNAs that would recognize several different sequences in a plasmid DNA, a circular piece of double-stranded DNA that we could purify in the laboratory. And what you're seeing on the slide here are marked in red 5 different sites in this plasmid DNA that were chosen as target sequences for single guide RNAs that we produced in the laboratory. And Martin did the experiment of taking those single guide RNAs, adding them to the purified Cas9 protein, and then incubating together with the plasmid DNA molecule in a laboratory test tube. And then to analyze the result of that experiment, he separated the cleaved DNA products in agarose gel system, which is shown here, which is simply a way of separating DNA of different sizes from, from each other. And What you can see in each lane of this gel system is that depending on where the guide RNA was directed to interact with the plasmid DNA, Cas9 would generate a cut, and then by also cutting the plasmid at a separate place so that two double-stranded breaks were introduced into the plasmid at one time, we could release these little fragments of DNA that I hope you can see on this gel system. That migrate at different positions representing different sizes of these DNA fragments. And I have to say that on the day that Martin Jinek did this experiment and got this result, we were just incredibly excited. It was the pure joy of discovery at recognizing that we not only understood how this bacterial enzyme Cas9 functions, but we had actually figured out how to engineer it as a simple two-component system for directing DNA double-stranded cutting. Now, why was that so exciting? Well, it was, of course, interesting to know that we could harness our knowledge in this way and engineer the protein to have this desired cleavage capability. But in addition, it also allowed us to imagine how CRISPR-Cas9 could actually be harnessed as a technology for something quite different in eukaryotic cells, namely cells like plant, animal, human cells, all of which treat double-strand breaks in DNA differently than the way they are treated in bacteria. And I'll show that on the next couple of slides. So in eukaryotic cells, when the cell receives or detects a double-stranded break to the genome, to the DNA in the cell, the break is detected and repaired typically before it can cause cell death. And the repair pathways involve either a non-homologous end joining event, shown on the left side of this slide, that can sometimes introduce a small disruption to the DNA sequence, or there can be an integration of DNA that has homology to the sequence of DNA flanking the double-stranded break. And in that case, a new piece of genetic information is incorporated into the genome at the site of the original break. This was research that had been done over the previous couple of decades before Emmanuelle and I did our work with Cas9, but we recognized that the activity of Cas9, its ability to introduce a double-stranded break to DNA at a desired position by directing it with these single guide RNAs, could allow scientists to introduce double-stranded breaks into a genome that could trigger the kind of repair that I'm showing here. And so we imagined a system working very much like in this video, where the Cas9 protein could be directed to enter the nucleus of a eukaryotic cell, and with its guide RNA, search the genome for a 20-base pair sequence that would match the 20 nucleotides, the 20 letters of the guide RNA that would provide a match. And when that match occurs, we now understand that Cas9 is able to unwind the DNA. It allows the Cas9 protein to generate a precise double-stranded break in the DNA by cleaving each strand of DNA. And then those broken ends are handed off to repair enzymes in eukaryotic cells that lead to DNA repair. And in the process of repairing the DNA, there is the ability to introduce a change to the genome at a precise place, which is really the definition of genome engineering. So, on the next few slides, I wanna share a few things that we've learned over the last few years about how Cas9 is able to achieve this kind of editing. In genomes of cells by triggering double-stranded breaks. And I'll start with showing a molecular model of the Cas9 protein, and this is based on a crystallographic structure that was solved originally by several different laboratories— our own, the lab of Martin Jinek, the lab of Osamu Noreki, and subsequently many others who have contributed to understanding the actual molecular basis for Cas9's function. And in this slide, what you can see is the white protein Cas9 holding on to its orange guide RNA and a blue double helical DNA molecule. And I'd like to point out that in this structure we can see the mechanism by which Cas9 uses an RNA guide to interact with DNA at a precise position, because you can see the orange and blue helix formed inside the Cas9 protein representing the interaction between the guide RNA and one strand of the target DNA molecule. This is the actual way that these proteins are able to find and hold onto DNA prior to DNA cleavage. The next thing that we learned in the lab in studying how this cleavage event actually works is that a series of experiments showed that Cas9 is a highly dynamic protein. It has to be to be able to handle DNA, unwind a double helix the way we know it does. It has to be able to move. And through a series of chemical experiments that allowed us to detect motions in different parts of the protein, we discovered that Cas9 undergoes a large conformational change as it holds onto DNA and catalyzes cutting. And I'll show that in this video here. So this shows the Cas9 protein alone morphing to the structure it forms when it binds to the guide RNA, which you can see now in orange. This is the, the structure that is able to search the cell looking for a matching sequence of DNA. And when that match occurs, there's an additional structural change in the protein that accommodates the DNA molecule so that you can now see the RNA-DNA helix forming inside of Cas9. And then finally, this yellow part of Cas9 swings into position so that it can cleave the DNA strand that is attached to the guide RNA. And this is a very important aspect of the chemistry of Cas9 catalyzed DNA cutting because it provides a mechanism for sensing the interaction between the guide RNA and the DNA and ensuring accuracy of Cas9's cutting mechanism. In work that was done just over the last few years, Fugo Zhang, a former postdoc in my laboratory and very sadly now De Syst did a series of very exciting structural experiments to reveal the shape of the Cas9 protein when it's engaged on a full-length DNA molecule. This shows again the guide RNA in orange, the DNA strands are in blue and magenta, and you can see how the DNA is held open by the Cas9 protein. And allowed to interact with the cleavage sites in the enzyme for precise double-stranded cutting of both strands of the DNA. And then in this image, you can see in green the, um, one of the cutting parts of the enzyme swinging into position so they can actually catalyze the chemistry required for DNA strand break. And then I wanted to also point out that in addition to these conformational changes that happen in the Cas9 protein itself, we also now understand that this protein is quite dynamic in the way that it interacts with long pieces of DNA. For example, chromosomes in cells. It has to be able to move very quickly along the length of DNA searching for a sequence that will bear a complementary match to the guide RNA. How does that work? It seemed like an extraordinary capability to us initially. And so in experiments that we did in the laboratory over several years with a number of former lab members and collaborators, we came up with the model shown here that I think is consistent with current data, suggesting that the Cas9 protein has the ability to bind and release DNA very quickly. And that allows it to search through very large, really vast stretches of DNA quite fast to allow interrogation of the sequence that's being searched for a match with the guide RNA. And so in this example, the guide RNA is in red, And you can see that our data suggests that when this Cas9 protein with its guide interacts with DNA, it begins to pry apart the two strands of the double helix to allow the protein to ask, is there complementarity for binding to the CRISPR RNA? And if there is, then our data suggests that the strands of the DNA continue to melt apart, that allows the RNA-DNA helix to form inside the Cas9 protein. And if that helix is perfect, or close to perfect, then the enzyme is triggered to cleave DNA. And this is quite an amazing mechanism and clearly allows bacteria to search the cell very quickly looking for viral DNAs to destroy. But in eukaryotic cells, this mechanism is equally effective at triggering double-stranded breaks that can be repaired and trigger changes in the format of genome editing that we now understand can be effectively catalyzed by Cas9. So, in the next part of the talk, I really wanna turn my attention to where this technology is going. And there's a lot that one could say here, so I'm really gonna hit on a few of the highlights. And first of all is the fact that genome editing extends across all of biology. It can be used for fundamental research, but also for exciting applications in public health, in agriculture, and in biomedicine. And it's very important also, I think, to point out that genome editing can be conducted in many different kinds of cells, and fundamentally in the two kinds of cells that I'm pointing out here. Category of cell is a somatic cell. That's a cell that is fully differentiated. It does not have the ability to create a new organism, versus a germ cell, which is a cell such as a sperm or an egg cell or cells in an early embryo that have pluripotency, and they're able to differentiate into many different cell types as an organism is forming. If genome edits are introduced in a somatic cell, those changes to DNA are not heritable. So they affect only one cell or one tissue type or one individual organism. But if genome edits are introduced into a germ cell, they have the potential to be heritable and to introduce changes that become part of not only an individual, but all of that individual's progeny. And this is, of course, very powerful when we think about using it in plants or using it to create better animal models of human disease, for example, as has been done using CRISPR-Cas9 in mice and rats. It's very different when we think about how it could be impactful in human biology and the, of course, enormous ethical and societal issues that are raised by the possibility of using germline editing in humans. And I won't say too much more about that, but there's a— it's become a very active area of my own work over the last few years, is to think about responsible use of CRISPR-Cas9, and in particular, its use in humans, and ensuring that there's transparency and careful thought that goes into work, particularly when applied in the human germline. But in somatic cell editing, I think there are extraordinary and exciting opportunities that will be developed in the near future, and one of them is in the area of correction to disease-causing mutations in humans. And this is just one of many examples, the opportunity to correct a mutation, well-defined, that causes sickle cell anemia, and can now be corrected using CRISPR-Cas9 technology. And this is no longer a potential opportunity, but it's actually been realized by using CRISPR-Cas9 to correct the, the disease-causing mutation in a patient and to show that this technology is safe and effective for this kind of treatment of genetic disease. I think this is a a very exciting way to imagine how this technology will impact human health in the future. And of course, with germline editing, one has to think about how this technology will advance in the future. And I certainly imagine that we will see increasing applications in germ cells, including in the human germline. And of course, that has to be managed very carefully, and I'm pleased that there's been an active international effort to control the use of CRISPR-Cas9, and certainly to encourage transparency, especially with the recent release of a report that discusses the science and the technology around human germline editing, and very importantly, establishes criteria for using CRISPR-Cas9 in the human germline in the future. In the last few minutes, I'd like to turn to where CRISPR technology is headed in the future. And one of the aspects of this work that's fascinating is the incredible diversity of CRISPR systems in nature. This continues to drive the field in terms of fundamental biology, understanding what these systems do in their natural settings in microbes, and of course, how they may be harnessed as technologies in other organisms, as was the case for CRISPR-Cas9. I wanted to mention briefly the effort to investigate new CRISPR-Cas systems, and one of the recent findings that we've had with our collaborator Jillian Banfield at Berkeley is the discovery that phage, bacteriophage, the viruses that bacteria use CRISPR to protect against, in fact, can also carry around their own CRISPR-Cas systems. And this is one example shown here, a protein that we named CRISPR-CasPhi that is entirely phage encoded. It's a tiny protein, but it nonetheless has the RNA-guided DNA cutting capabilities that we discovered originally in CRISPR-Cas9. And so this is, I think, a fascinating example of nature's diversity, as well as opportunities possibilities that this protein might provide for future applications, including in cases where one could benefit from having a very tiny protein, a small gene that encodes that protein, that could be potentially more easily delivered into eukaryotic cells. We've also been very interested in other biochemical activities of CRISPR-Cas proteins. And I'll mention here very briefly that in research done originally by Alexandra East Silecki, a former graduate student in my laboratory, and Mitch O'Connell working in partnership with her, they discovered that a class of CRISPR proteins called Cas13, which are naturally RNA-targeting enzymes, have a biochemical activity that could be harnessed as a detection mechanism. And this was an experiment done originally by Alexandra E. Silecki and actually suggested by Jamie Cate, the idea of putting fluorophores onto small pieces of RNA that were cleaved upon Cas13's detection of an RNA sequence using its RNA guide. And this system turned out to be highly effective at detection of RNA molecules all the way down to about picomolar levels. This was in just very early experiments with an off-the-shelf kit called RNase Alert. In the laboratory, but it also triggered our interest in the biochemical activities of other families of CRISPR-Cas proteins. And in research done a couple of years ago, Janice Chen, a student in my lab at the time, showed that another family of enzymes called Cas12 also have the ability to cleave single-stranded molecules, in this case single-stranded DNA, upon recognition of a double-stranded DNA target. And in this experiment, what Janice Chen showed is that the Cas12 protein, upon recognition of a double-stranded DNA, is able to cleave single-stranded molecules of DNA, shown in this experiment in the panel in the center of the slide, showing very rapid degradation of a circular single-stranded DNA molecule in this experiment, which is an activity, a biochemical activity, that we do not detect for Cas9. And working with a colleague, Joel Polevsky, at University of California San Francisco, we were able to use this Cas12 activity for detection of the human papillomavirus in human patient samples, and even to distinguish between two different strains of the human papillomavirus, which is what's shown in the panel in the center of the slide. And so this told us that not only could CRISPR-Cas enzymes be useful for detection, but they could also be useful for specificity in figuring out what type of viral signal might be present in a patient sample. And of course, in the current SARS-CoV-2 pandemic that has become a worldwide human health issue, we and others have been using this type of CRISPR-Cas detection to identify samples in patient saliva or in patient nasal swabs, and to do so very quickly. Ultimately, we hope with a readout that will involve a simple mechanism with a cell phone that can record the results of these diagnostic tests and help people everywhere to screen themselves against this virus, and importantly, provide a mechanism for future pandemic preparedness because of the programmability of the CRISPR-Cas enzymes, which is a fundamental property of their biology and their ability to be harnessed as technology. So I'd like to close by just pointing out that the RNA-guided gene regulation that we observe in CRISPR-Cas systems is fundamental to the way they work in bacteria, but also to the way they operate as technologies for genome editing and beyond. Delivery and control are going to be key in the future, so we need to have better ways to deliver the CRISPR-Cas9 and related proteins into cell types of interest for genome editing, including into human patients. And fundamental research will continue to drive the field forward as it has in the past and as it will in the future. The possibilities are endless, and I'm excited about what will happen in the future, both with fundamental research and with the applications that will solve real-world problems in human health and the environment. I'd like to thank my past and current lab members with whom I've had the joy of doing science together over many years. I wanna thank the many colleagues who have been involved in the fields of genome engineering, in DNA repair, and in the applications of genome editing that have made the field so exciting over the last few years. And finally, I wanna thank my colleagues at the University of California, Berkeley, who I've had the joy of working with, and where we share together the dedication to public education and research that makes our work so rewarding. Thank you very much.