Jennifer A. Doudna, PhD | UCLA School of Medicine 56th Annual Lectureship

all right well welcome everyone to the 56th annual lectureship of the David Geffen School of Medicine I'm John Massey OTT I'm the Vice Chancellor of Health Sciences and we're honored to have you here today to hear our lectureship Award recipient Jennifer Doudna dr. Donna is a Howard Hughes Medical Investigator she's professor of chemistry and molecular and cellular biology at UC Berkeley and she also holds the lee kai-shing Chancellor's chair and biomedical sciences dr. dadno welcome to UCLA we're excited to hear about your amazing work that's changing not only biology but probably society the annual lectureship is a long-standing tradition of the School of Medicine that reinforces the importance of scientific exchange and breakthrough discoveries in biomedical science the essence of a great university is discovering new ideas new knowledge all of which benefits society innovators like dr. Doudna embody all these goals and are the quintessential quintessential expression of human curiosity so your program says that interim dean Kelsey Martin is going to introduce dr. dad no but we change that and I'd like to ask Doug black professor of microbiology immunology and molecular genetics and expert on RNA biology to introduce our speaker Thank You Dean Massey ODA well I feel like I'm introducing I don't know Bill Gates or something there so it's really my pleasure great pleasure I think to introduce my long term colleague jennifer doudna for the Geffen School lectureship Jennifer really I think she can only be described as one of the most admired and influential scientists in RNA biology and I think I'm going to skip the long list of Academy elections and breakthrough prizes and so on and try to say something about her science and and really some of the science that that led up to what she's doing now which which all of you are most familiar with so Jennifer actually is from Hawaii she grew up in Hawaii and came to stateside to go to college he went to Pomona College close to here and then went on to do her PhD at Harvard Medical School and Jack szostak slab and at Harvard she really pioneered the entry of the szostak lab into ribozyme ology and the study of the early RNA world and I think that's where she got her start in RNA biology and interest and that's stuck with her and she went on then to do her postdoc with Tom Cech at Colorado studying the structure and enzymology of RNA again and she took this work with her when she started her own lab at Yale in 1994 where soon thereafter she determined and published the structure of the p4 p6 domain of a group one intron and this result it seems long ago now but it's hard to really overstate its impact on the RNA field it was the largest RNA structure that had been solved today quite a bit larger than the long study tRNAs and within this structure it illuminated a number of structural features and principles of RNA folding that are really now commonly held principles in the study of RNA function and in folding so Jennifer moved to Berkeley in 2002 where her positions were just described by the Dean and at Berkeley she has been continuing her studies of ribozymes and RNA biology and biochemistry but she's also branched out into many other areas of RNA biology she's she's worked on areas of translational control viral iris function enzymology of the RNA interference pathway just to name some of them were more larger projects and in through this she's trained many some scientists who are now leaders in all of those fields and so if I were to try to describe you know Jennifer's kind of special style or traits of science I would say that through all of this work she's displayed really just a remarkable skill for seeing what the key information that's needed to drive a field forward to what experimental approach often grounded in structural analysis will really break you know change the question that that's being asked I think it's a rare talent that the only other person I think I've with this talent is some Phil sharp and so a number of years ago Jennifer became interested in the mechanisms of host defense systems in bacteria and in particular the the enzymes and and gene products that were encoded in what were called CRISPR lo Sai that mediated bacterial resistance to infection and so what's ensued from those studies I think is really one of the great demonstrations of how basic research into interesting but maybe not particularly mainstream biological questions can have an absolutely pervasive impact on on much larger fields of science in medicine as all of us are now aware of some of the things that have come out of those of these projects and so to give you just a small amount of the history of these projects without too much about the biology of crispers and how and their diversity the Dowden lab first started studying what are called the type 1 CRISPR systems publishing a beautiful structure of the multi subunit cascade complex that mediates DNA recognition and cleavage of the invading DNA and then a subsequent to that or as time went on they turned their attention to what are called type 2 CRISPR systems where in place of the Cascade genes there was a special gene called caste 9 and in studying this there's these type two systems the doubt in a lab in collaboration with Martin eunuch and and Emmanuelle Charpentier in Berlin developed biochemical assays for this this casts nine polypeptide and showed that this was a single polypeptide nucleus that could be programmed to cleave a wide variety of sequences through its interaction with a crisper RNA and another RNAi cofactor and so that paper I always like to recommend this paper to students that was published in August of 2012 it's really a landmark result of nucleic acid biochemistry and it's really changed everything that we do as molecular biologists and it's was followed in January 2013 by results from three labs that the Doudna lab Fung John's lab and George church's lab showing that one could in fact Express the cast nine protein in cells along with the proper RNA and achieve highly specific targeted cleavage of a mammalian genome at sites of one's own choosing and so that the efficiency and ease with which this this worked immediately caught people's attention as a tool for the the developing field of genome editing and it opened really the floodgates of work on this on this enzyme by by many labs and I think in 2013 so that that that paper from those three labs those three papers were published in January 2013 in the rest of 2013 there were more than a hundred papers published showing targeted cleavage in really all variety of organisms and developing the CRISPR cast nine into really a wide variety of incredible tools I was actually lucky enough to be on sabbatical in Berkeley at the end of that year which was really just a remarkable experience to see this the pace of progress the number of labs every day there were results coming out and that rate of element has really only increased since then it's and I think crap I think we can say that crisper cast nine is is a standard tool for all of us and in modern biology and so Jennifer's lab went on I mean they've they've been major contributors to the tool development but I think it's all they've also maintained their focus on basic mechanisms and the last year or so they've published lovely papers on the type three CRISPR systems on how on how cell bacterial cells acquire the targeting sequences and other quite fascinating aspects of CRISPR biology and so finally then III think I need to comment that that powerful new technologies you know raise questions for society and and their appropriate use and Jennifer has been at the forefront of calling for public discussion of how these tools should be used what's an appropriate use of this technology how genome editing might be by use usefully and how perhaps it shouldn't be used and so these kinds of discussions these public discussions were reached a milestone last week at a remarkable meeting at the National Academy where many of the stakeholders and interested parties met a meeting that Jennifer was really instrumental in organizing so you can see then that there's much for Jennifer to tell us about and before I bring her up here to talk about her talk the deans are going to come up Dean Massey Oda and Anne Martin is that now yeah you weren't moving I was and Jennifer 56th annual lecture a ship award from the School of Medicine there you go the fake check reduction I really appreciate that and it's been a great day here and I think you know it's always a great pleasure to see a lot of students here because I think that this work as I'll tell you about today and you heard a little bit about in the introduction is a great example of how your own curiosity about nature and how biology works can lead you in unexpected directions and I have to say over the course of my career so far that's really been my experience is just by following what I thought was interesting and of course what my students thought was interesting we it sort of led us in directions that I never could have anticipated including the story that I'll tell you about today so so what I thought I would do today is really tell you about the origins of the CRISPR biology research in our own lab and how it led us to thinking about this as a technology for genome engineering and then I want to tell you a bit about how this actually works because we're really a lab that focuses on molecules and mechanisms we like to understand how these kinds of activities occur in cells and then I'll tell you a bit about where I think this is going in terms of its application in the clinic and then at the end I want to say come back to this point that Doug made at the end about the ethics of genome editing and where this is going in the future so I want to really start by just just pointing out that you know if just even a few years ago if you had thought about the the ability to change the DNA sequence in a Cell in a very precise fashion it was still somewhat in the realm of science fiction and in fact I can remember being a graduate student in Boston in the 1980s when Peter Dervin and other others were actually already thinking about you know how you could actually do this in the context of the genome of an entire cell and the challenge really was could you envision a way to make a very precise change to the DNA sequence so precise that you might be a to change a single base pair imagine that just a few atoms in the entire genome of say a human cell and be able to fix a mutation that would otherwise cause disease sort of an amazing thing to think about sort of very analogous to you know being able to make a very precise change to a document in a word processing program and and for us the you know the path to this kind of technology actually came about through a through our effort to understand how bacteria fight the flu and I noticed that my resolution is cutting off the bottom of the slide should I try to change that or not I don't know it might be risky so I'll read you the bottom of the slides but basically it was curiosity about how bacteria fight viral infection and so I want to tell you I'm gonna sort of do three things today I want to talk about how we got from studying up sort of as Doug said kind of a somewhat obscure area of biology to a to a biotechnology and then I want to tell you a bit about how this technology actually works from the standpoint of the molecules involved and then at the end I'm going to come back to this question of how we think about using this technology both in the clinic and also from a bio ethical and societal perspective so for us you know the the CRISPR systems really came first to my attention about ten years ago when a colleague of mine whose picture is unfortunately not showing at the bottom of the slide a bajillion banfield who's a very well-known scientist at UC Berkeley called me up one day and you know her lab does met a genomic sequencing of bacterial genomes and she contacted me because she said you know we're seeing a very unusual sequence signature in the genomes of many of the bacteria that we're investigating and mostly these were bacteria that have never been cultured never even been identified and in her research they were coming across not only these bacterial genomic sequences but also the the viruses that infect these organisms and what was very very interesting was that a number of these bugs had a repetitive DNA sequence in the genome that consisted of a sequence of DNA typically 40 base pairs or so in length that was repeated over and over in the genome and and what was very interesting was that in between these DNA repeats were sequences that were derived from viruses and again these sequences were 30 to 40 base pairs in length and what's not showing here is three papers that were published in 2005 from three different bioinformatics groups that all pointed out that this sort of sequence signature had the sort of had the characteristics of some kind of a system and bacteria these were conserved and the fact that bits of viral sequence were being integrated into the bacterial genome you know sort of made people start to wonder if this was some kind of acquired immune system in bacteria and and what was also interesting was that adjacent in the genome were typically CRISPR associate ER cast genes that seemed to be Co evolving with these these DNA sequences so it looked like some kind of a system and these had come to be called a clusters of regularly interspaced short palindromic repeats or crispers so if you see that in the media it's actually referring to these sequences right here these sort of signatures of what emerged and came to be known and understood as a bacterial acquired immune system so why did you'll call me while she called me because as Doug mentioned you know we've had a long-standing interest in my lab in understanding how RNA molecules are used in cells to control the expression of information in the genome and one hypothesis at the time was that maybe these sequences were actually operating at the level of RNA in other words the cell might be making an RNA copy of these sequences and then using those RNA molecules to help find matching viral sequences much the way eukaryotic cells use a process called RNA interference so we began sort of playing around with these systems and what emerged over the next several years and in a sort of another interesting twist about the biology here it was really ashlee genetic experiments done at a yogurt company the misko that showed that that these systems actually do operate as an adaptive immune system so they allow cells to detect viral DNA that is gets into the cell I'm showing it here as a phage injecting its DNA bits of that DNA can be integrated into the CRISPR locus and that integration occurs such that each newly integrated bit of sequence is flanked by a copy of the DNA repeat and we in work that I won't tell you about today we've actually in the lab done a number of experiments to understand exactly how that integration process works and then these sequences are transcribed into RNA and those RNAs are then processed into shorter bits that each include one bit of sequence from a virus they assemble with proteins encoded by the cast genes to form RNA protein targeting complexes that then use the information in the RNA to base pair with matching DNA sequences and allow these proteins to cut up the viral DNA so it's a very very nice way that bacteria create a what's effectively a genetic vaccination card in the genome that creates a permanent record of viruses they've been exposed to over time and then uses that genetic information to protect the cell so we started really focused initially on this central part of the pathway namely how these RNAs are made and how they form RNA protein targeting complexes that can interact with DNA and so and as Doug mentioned we started off working on what are now called the type 1 CRISPR systems and involve multiple proteins that are part of these targeting complexes and then I in 2011 I went to a meeting of the American Society of microbiology and I met Emmanuel sharp NTA who is a medical microbiologist working on streptococcus pyogenes an important human pathogen which has a different kind of CRISPR system in the genome namely a system that had only one gene called Castine that had been shown at that time genetically to be important for the this acquired immunity in that bug and so we decided to team up to figure out the function of that the protein encoded by the caste 9 gene and so that project led to research that was done in both of our laboratories by Martin genic in my lab and by Christophe Chomsky in a Manuel's lab and what these two guys figured out was that caste 9 is a protein shown here in this blue that binds to 2 double stranded DNA opens it up sort of locally and is able to make a blunt double-stranded DNA break at a site determined by a xx nucleotide sequence in an RNA molecule that is derived from the CRISPR locus so one of those transcripts from the CRISPR sequence in the genome and so here's the the targeting sequence right here and importantly this is a system that requires a second RNA to be functional so in bacteria this is an RNA called tracer that is important not only for generating the mature CRISPR transcripts these molecules but it actually stays associated with the end of the crisper RNA where it forms a structure that can bind to the caste 9 protein so you really have to have both of those RNAs to assemble a functional targeting complex and we also figured out that these these targeting sites have to occur next door to a little motif in the DNA that for this protein is a GG dinucleotide motif and when that's all there you have the rnas and this little motif and complementarity to the guide RNA then you get this reaction and so it was through that biochemical dissection of how this actually worked that led Martin genic in the lab to figure out that he could trim away some of the extraneous or at least functionally extraneous sequences from these two RNA molecules and he figured out that he could actually create a simpler system than what nature has done by turning this into a single guide RNA by connecting the ends of these molecules to form a complex that looks like this where we have in a single transcript both the targeting information and the caste 9 binding information and so in in a very sort of simple experiment that Martin did to test whether the single guide this single guide type of construct could actually work for DNA targeting by cast 9 we did an experiment and very sadly the cool result is cut off the bottom but but I'll tell you what it is and that is that so we had a DNA circular DNA molecule a plasmid and Martin designed five different versions of the single guide RNA that were adjacent to these GG die nucleotides and then just incubated this purified DNA with the Castine protein programmed with one of these different single guide RNAs together with a different restriction enzyme that cuts about 60 base pairs upstream and when you look at this the result here what you could see was that each little piece of DNA that was released and this is actually an agarose gel system each little piece of DNA released was exactly the size that would was predicted based on where cast lines should be cutting the DNA guided by these different constructs so we really this was for us really the moment when we knew we had a two component system that was quite simple to use where you could easily manipulate the site of DNA cleavage by just changing the sequence of RNA in this in this single guide type of construct and just to show you a little bit in a little more detail how this actually works so this is a 3d printed model of the Cassadine protein that's based on a crystal structure that Martin gen-x lab solved and published last year he's now at the University of Zurich and what this shows you is the protein in white it's got a sort of a cleft running down the center where this guiding RNA and orange is located and the way this works is that the RNA actually makes a helical interaction with one side one strand of the double helical DNA molecule as it traverses through the proteins you can see the rna-dna hybrid right here the other strand of DNA is opened up and then the protein has to molecular blades to active sites that come in and actually cut the DNA so it's a very precise kind of almost like a molecular scalpel to cut DNA and so you might be thinking well that's fine but how does that get us to genome engineering technology this is where I have to to bring in all of the great work that had been going on really for the past several decades in which many labs around the world had come to appreciate that in animal and plant cells there's a very robust mechanism for detecting double-stranded DNA breaks that occur in the genome so if this is the genomic DNA and there's a double-stranded break that occurs that break can be repaired by a couple of by different pathways the primary ones are non-homologous end joining in which the DNA is ligated back together typically with a little insertion or deletion that the site of repair or if there's a donor DNA molecule present in the cell then this can actually be recombined into the DNA to repair this break with the insertion of new genetic information at that site and so it had been appreciated that if you could control where double-stranded breaks were induced in a genome you could actually trigger these pathways and thereby trigger very precise changes to be made to the DNA of a cell and so you're probably many of you here are familiar with some of these protein based technologies that had evolved to do this zinc finger nucleases and talons among the most famous of these as well as homing endonucleases and these are proteins that can be designed or programmed to have very specific DNA binding activity and then by coupling them to a DNA cutting domain you can actually have a site specific DNA cleaver and these can work extremely well for inducing precise changes in a genome of cells so lots of excitement about these technologies but they haven't really taken off very broadly because it involved using them involved pretty significant protein engineering and you had to make a new protein for each experiment in a Cell and so we looked at this and said well you know if we have a single protein who's who's whose DNA cleaving activity and specificity can be controlled by simply changing a short guide RNA sequence that could be a very nice way to a very simple way to do this where you don't have to change the nature of the protein in every experiment but simply change the guiding RNA and so that's really what we proposed in this original work that we published with the manuals lab I'd like to show you a little movie that just illustrates how we kind of imagined that this might work inside the nucleus of an animal or a plant cell where the DNA is of course highly packaged and this just shows zooming into the cell and you know that that in eukaryotic cells the DNA is in the form of chromatin so the DNA is wound around nucleosome core proteins histone proteins to form these these structures that you can see here in green and so somehow this bacterial enzyme has to search through the DNA that's packaged like this in these cells to find a in principle a single site that has a sequence that matches the sequence of the guide RNA that it's programmed with and when it finds that site it unwinds the DNA it binds as I showed you before with it forming this rna-dna hybrid and then the DNA is cut and and somehow released to repair enzymes in the cell that can then repair this and in this example by actually recombining in a piece of DNA that it results in the integration of new genetic information at the site of the break and so this turns out to be a way that one can actually change the DNA sequence in a very precise fashion and for reasons that we're still working to understand this bacterial enzyme seems to be quite capable of dealing with the kind of higher order structure that it finds in a eukaryotic cell and this seems to work very broadly across different cell types and tissues and and whole organisms so let me now tell you that turn a little bit to how we think this actually works and so my lab has continued to investigate you know this sort of molecular process with the hopes of understanding I sort of satisfying our basic curiosity about this but also improving it as a technology and and turning it into something that we hope eventually will do things like allow the cure of genetic disease and so people ask you know sort of why why did this sort of take off very quickly the way it did and I really like to point out three things one is the power of base pairing and I think you know biology over and over has showed us that you know if you look at the RNA interference pathway in eukaryotic cells for example that takes advantage of RNA in that case RNA RNA a base pairing for specificity here we have rna-dna hybrid ation that is being utilized for specificity and it's just a it's a powerful and relatively simple way to to program a protein and there's many applications of this of course and as I'm going to show you this is a system that has really evolved to be highly very fast and and really quite accurate in terms of its targeting capabilities because of course in bacteria it's it's sort of a life-or-death selection so it really has to be good at finding and in allowing cells to get rid of viral DNA sequences so so this is a protein that a couple other things I want to just point out so first of all the cast line protein is quite easily modified so many labs now have been able to make different versions of this protein so I think it was Stanley Chi at Stanford working with Jonathan Weitzman and Wendell Lim that first made the what they called e cast nine sort of the deactivated form of this protein that doesn't cut DNA but allows transcriptional control and is still in a sort of an RNA programmed way so that's a an application of this that allows not permanent changes rather than making permanent changes in the genome you can actually control the way information is expressed at particular genes and then of course many different versions of what I call chimeric cow sign so this means linking cast line up to other enzymatic activities that allow modification of DNA in particular ways and there's lots of ongoing work doing this kind of thing also I want to point out that it's naturally multiplex so in bacteria this protein is programmed with multiple different guide RNAs to allow protection against different viruses much the way our own immune systems work and so of course scientists can take advantage of that by programming this protein with multiple different guide RNAs in to affect changes in the genome at multiple sites in the same experiment something that was really not possible with previous technologies and as I'm going to show you now it's really a protein that has remarkable ability to search the genome very quickly and find targets really quite accurately so one of the first experiments that we did to start investigating how this protein actually works was involved a great collaboration between our lab and the lab of Eric green at Columbia so Eric had come to Berkeley and he talked about his work using what he calls DNA curtains these are single molecules of DNA and there he was using whole phage genomes in these experiments to look at how enzymes search through a large DNA molecule to do things like repair a mutation and Sam Sternberg a student in my lab at the time saw this talk and said this would be such a great system to use to study these CRISPR proteins and so we got together with Eric and in in a series of experiments done by Sam in my lab inside reading in Green's lab they did experiments using the Curtin system to investigate the way cast nine searches through a large DNA molecule and I just gonna summarize what the findings of this and show you a little bit of what these experiments look like so this is actually showing you a movie that illustrates the way these DNA curtains are set up so each of these green strands is a 48 KB DNA molecule corresponding to phage lambda DNA and these are molecules that are tethered on one end to a slide and then when we have buffer flowing across the slide these molecules are extended and we can visualize them and also visualize the way that proteins that are labeled in this case with a quantum dot can interact with the DNA so we have cast nine that's labeled with a quantum dot and programmed with a particular guide RNA in this experiment you can see that a lot of those proteins are lining up on the DNA at a place that corresponds to the site that is recognized by this particular guide and so by using that kind of a strategy we could investigate what happens when you program this protein with a mismatched guide for example and how long does it take to find a target and things like that and what came out of these experiments really three things one is that we found that dna-binding really begins at these Pam motifs so rather than this protein searching initially for a 20 base pair match to the guide RNA it really looks first for these sequences and only then does it interrogate the adjacent DNA for a match we also figured out that binding to the pam triggers a change in the protein structure that leads to DNA unwinding and we didn't initially know about the protein structural change but we could certainly tell that the DNA was opening up after this Pam contact occurred and only then was it in a competent state to cut the DNA and finally we found that then sorry I cut off the bottom here but basically we found that this protein stays very tightly associated to DNA even after the DNA is cut so very high affinity binding to the product and so Sam Sternberg really wanted to investigate sort of this question was sort of hinted at by these single molecule experiments because what we found was that cassadine was spending more time on portions of this DNA that were more G rich and so it really suggested that it might be spending time there at those regions because of these pam motifs and interactions with this site and so to test that more directly sounded a biochemical experiment in which we took a DNA and this is sort of just illustrating how we often do these experiments in the lab is we take a DNA molecule it can either be 2 DNA all ago nucleotides annealed together it can be a larger piece of DNA that has a sequence with a with a pam site in it that matches the sequence of a guide RNA molecule which is shown down here and when this base pairing occurs between the RNA and the DNA the DNA opens up and then the protein can actually cut this DNA sequence and so what Sam did was to set up an experiment like this a biochemical experiment in which we wanted to investigate the ability of different types of DNA molecules to compete for binding to the caste 9 protein and so the idea was to take DNA substrate that we could radio label as a target site in it here's the Pam and so this is a very efficiently cleaved by the cast line protein when it has a guide RNA that matches this sequence and the question was could we add different kinds of unlabeled competitor DNA's shown over here that had absolutely no complementarity to this target site but instead had just different numbers of Pam's and the question was could any of these compete for binding to cast line and thereby reduce its ability to bind and and cut this this actual substrate and the result of that experiment is plotted here and what we're basically plotting is the cleavage rate as a function of concentration of the competitor DNA molecule over here and what you what you're looking at is basically the fact that the more Pam's we had in the DNA sequence the greater the ability of that competitor DNA to prevent binding to the substrate by just by competition which really suggested that that this is a protein that really is interacting largely with these pam motifs rather than with an initial recognition of this target site and really to test that idea more directly Sam made a competitor that looked like this where we had an exact match to the target site in the substrate DNA and a single mutation in the Pam site of this competitor DNA molecule and this molecule is as lousy a competitor as this one here so in other words it's really about the Pam rather than this target site that in allows this DNA molecule to interact with the CAF's line protein so of course we wondered you know how does this work and what happens when when this Pam interaction occurs and one of the things that emerged out of a collaboration with my colleague Evan nogales at Berkley and involved two students that you can't see the names of here Sam Sternberg and David Taylor was using negative stain electron microscopy to visualize the Castine protein as it went from the structure of the protein alone to the surveillance complex where it's bound to the guide RNA and then finally to a complex that's assembled with a target DNA molecule even though these reconstructions were at very modest resolution about 30 angstroms resolution we could actually see very clearly in these e/m images that the protein underwent what looked like a quite a significant change in conformation upon binding to the guide RNA that opened up a channel in the protein where the rna-dna hybrid would form once this complex assembled with a substrate and I just want to show you a movie because what's been really sort of fun then over the last couple of years is as crystal structures of cast line in different states of assembly have become available from us and in the lab of Osamu Nawrocki and also from martin genic it's been possible to create a movie that morphs between these different structural states and you'll see something really quite remarkable in this movie and I should say it was made by a rotation student in my lab graduate student band LaFrance and so this movie starts off with cast 9 in just in the sort of the state where it's the protein alone no guide RNA bound and so you'll start off by seeing this here's the protein structure and as it morphs to the structure bound to the guide you saw a huge change rotation and conformation of this part of the protein right here that opens up a channel where the guide strand of the RNA and orange ends up located serve in the center of the protein and then as it morphs to the structure bound to a strand of DNA you can see there's an additional change in conformation of the protein here to accommodate that are the formation of this rna-dna hybrid and then this final conformational change was initially purely our imagination because this is actually one of the catalytic domains of the protein one of the cleavers of DNA and in all of the available crystal structures up until that point this domain was in the wrong place to actually cut the DNA so we knew there had to be an additional conformational change of this protein to allow it to actually cleave DNA and in work that I'm not going to show you in any till today but we published very recently with Sam Sternberg as the first author Sam and other students in the lab were able to show that this actually we can actually detect this conformational change and the others as well by putting pairs of dyes on the surface of the protein that allow detection of conformational changes by looking at fluorescence resonance energy transfer between those dyes so we can really map out these conformational changes through that sort of a biochemical assay and so this is really just a cartoon that illustrates summarizes our model for how Cass and we think caste 9 interacts with DNA potentially even DNA that's involved in chromatin based on up until this point in vitro experiments doing experiments with the single molecules as well as in bulk biochemical assays and and our data really pointed to a model in which the protein RNA complex rather than binding on one end of DNA and sliding to find a target instead had very rapid binding kinetics or binding and releasing DNA very quickly slowing down when it encountered a pam sequence which allowed time for interrogation of the adjacent DNA sequence and only then if there was complementarity between this sequence of DNA and the guide RNA would this interaction be stabilized sufficiently that the DNA could begin to unwind the rna-dna hybrid could form and then eventually that final conformational change could happen that would allow DNA to be cut but this was all based on you know in vitro data and we kept thinking wouldn't it be great if we could actually see how this works in a in a real cell and so this was the challenge that was taken up by Spencer Knight a chemistry student in the lab who joined my lab and the lab of Bob Teigen to use super-resolution microscopy to try to investigate the search mechanism of caffeine as it moves around the nucleus of live self and his idea was to explore the way that caffeine might be able to interact with DNA that was not just you know naked on a slide but really was wrapped up in the kind of higher order structure that we know occur in chromatin and and so the question was really you know could we understand not only how it deals with chromatin structure but also how it finds sites in genomes that are much larger than a bacterial genome and how it deals with higher order nuclear organization so the way that DNA might be located in sort of neighborhoods inside the nucleus and I just want to summarize a couple of things that Spencer did so he figured out a way to label cast nine in living cells and this is done by making a fusion protein so we were actually using the the deactivated form of the enzyme so it doesn't actually cut DNA so which allows us to visualize it more easily and we hook it up to a what's called a halo tag this is a little a protein domain that has been modified so that it will react sort of permanently with a small molecule that can be a fluorophore like this in living cells and so this can actually be introduced into into live cells that are expressing this fusion protein and there's then this very specific chemistry occurs that puts the fluorescent tag on this fusion protein and so I'm going to show you a couple of little movies that show you fluorescent dots moving around in cells and that's actually how we're doing it it's using this kind of a tagged version of cast line and so one of the first things that we wanted that Spencer want to do is really just to understand how the behavior of cast 9 what it looks like when it's inside a living cell and then how it changes when we program it with different kinds of guide RNAs and so I want to first show you a movie that illustrates this sort of looks is what the some of the raw data looks like we're taking 10 millisecond snapshots of the nucleus of a living cell that contains this fluorescently labeled cast line you can see these little dots and kind of moving around so these are individual particles of cast 9 and one thing that we noticed right away was that these particles are moving incredibly fast so they're very very very fast kinetics what I'm plotting over here is is plot of the log of the diffusion coefficient of these particles as a function of frequency of particles with this behavior and you can see that we have for both the protein alone the APO form of Kassel and the protein programmed with what Spencer calls a nonsense guide so it doesn't recognize any particular site in the genome these both have very similar behaviors very rapid diffusion around the cell and much fat moving much faster than what we see for a histone protein for example h2 B which is one of the proteins that is responsible for the higher order chromatin structure so that was one sort of interesting observation and then when we program cast line with a guide RNA that recognizes about 300,000 sites in the genome so it should be allowing caste 9 to park itself at many different places in the genome we actually see a very dramatic change where now many of these particles have much slower kinetic behavior in the cells and you can sort of see that here a lot of these particles are kind of hanging around for quite a long time at a particular position and we think that's because they've found a target site in the genome and they just hold on and and you can almost see biphasic behavior we still have some particles moving very quickly but we have a large population that have slowed way down and so this has allowed us to start asking questions like how long does caste lines stick around on a site that has a perfect match to the guide RNA versus sites that have one or a few mismatches with the guide and what we find is that really the lifetime of these particles on a bona fide target site is minutes in these experiments versus less than a second for sequences that have mismatches with the guide so big differential there finally I wanted to show you another experiment that he did so we want wondered about access to different types of chromatin so you probably know that in cells you know we have euchromatic regions that are sort of more open and active versus heterochromatic regions that are highly compacted and so what what Spencer did in this experiment was he had a fluorescently labeled protein called hp1 which is involved in heterochromatin that is allow us allows us to figure out where heterochromatic regions are in the nucleus and then we can on that sort of background we can track these fluorescently labeled cast line particles as they move around this labeled DNA it doesn't look like much when I show it to you like this but when we stack up these images is what we see is that so here are the these very bright regions that are marking heterochromatic parts of the nuclear DNA and you can see that when you look at these tracts of cast 9 we see that that although most of the tracks are in the euchromatic parts of the nucleus we do see occasional forays into heterochromatic regions and that's actually consistent with the fact that we see evidence for gene editing in those types of those those regions it's sort of remarkable that somehow this protein is able to deal with DNA even in these compacted regions and I think that might actually be just saying something very interesting about chromatin dynamics that we're still trying to figure out with respect to caste line and I want to show you one little sort of recent observation that we have so you may know that in nature there are multiple different types of caste line proteins and you know people have been interested in this because they vary in size so we've been discussing up until now in this talk the this the protein called from strep pyogenes which is a type 2 a protein which has a molecular weight of about 160 kilo daltons and there are some much smaller proteins belonging to the type ii c cast lines that are about 25% smaller or so and those have been interesting from the perspective of delivery because they might be easier to package into viral vectors and things like that so you know one idea was well you know if they all work the same why don't we just pick a smaller version that Nature has evolved and so we've been studying some of these different caste lines in the laboratory and using biochemical methods to investigate their activities and one of the things that emerged early on not just from us but from other labs seongjong and George church and among others is that these type 2 C proteins are actually not typically not very robust at genome engineering and so why is that and I just want to show you one observation that I think is intriguing and that is a comparison that we did of a of this strep pyogenes caste 9 that we've been talking about cutting DNA versus one of the type 2 C proteins that we call we abbreviate CDI and so the CDI protein is not has not been observed to catalyze genome editing in in cells and the question was why not because it seems to be active in in in vitro but when we did a careful biochemical analysis this is plotting a fraction of DNA cleaved over time what we saw that was that whereas the pyogenes protein is very fast at cutting double-stranded DNA so this is doing an experiment where we have a double-stranded DNA substrate like this the CDI a smaller a Cassadine protein is very slow right so it could cut DNA but it's quite the kinetics are very different and so we started thinking about this and wondering if part of the issue here so when we compare a structural sort of predicted structural model of this smaller caste 9 to the structures of this bigger protein we noticed that part of the parts that are missing and the smaller protein correspond to parts that we think are actually involved in the helicase activity the unwind ace activity of this protein and to test that possibility what Mitch O'Connell and Anne boma in the lab did was to generate a version of this DNA substrate that has exactly the same sequence on the bottom strand so it can still make the full 20 base pair match matched rna-dna hybrid with cast line with its guide RNA but had mismatches over here in the DNA right next to the door to this Pam motif so the idea was to use a sequence of DNA like this where we've just sort of destabilized these first two base pairs thinking that we might give the protein a hand with opening up the DNA when we do that now what we find is that when we test this CDI protein we find that it's now got very very fast kinetics cutting this DNA substrate and so we think that this really suggests that this is a protein that has a reduced ability to open up DNA which may explain why it's not so great at genome engineering and when I one thing that we're currently thinking about is how we can take the scaffold of these smaller enzymes and turn them back into robust HeLa cases by introducing some of the features that are found in these larger proteins and thereby maybe making a smaller version of cassadine that will be robust technology for DNA for genome engineering so I just want to say a little bit about this so one of the things we're thinking about a lot in the lab in terms of applications is what the challenges are going forward so now that there are robust tools like this to to cut DNA and induce changes in cells that's great for research purposes but what if we want to use that as a therapeutic and I think really from my perspective the the challenges going forward are really about delivery how do we get it into cells in especially in a tissue specific fashion how do we control the way that DNA is repaired after it's cut is clearly a challenge we want to be able to control that in the future if we really want to have precise accurate editing that occurs the same way in every cell and then of course thinking about the societal and ethical issues that come up when we edit certain types of cells like thinking about editing human germ cells germline cells like sperm or eggs or embryos and so I just want to briefly mention work that we've been doing with colleagues at UC San Francisco in which we took a sort of to sort of think about how do we address this question and of course we're biochemists so we approach this from the perspective of thinking about the way that we work with this protein in the laboratory and so many labs have used these strategies here to introduce the cast line protein and its guide RNA into cells namely encoding it on a DNA molecule or encoding it in in the form of RNA a messenger RNA that would encode the cast line protein together with the guide RNA and those can work obviously very well when we introduce those into cells it's harder to think about how we get these molecules into into into an organism in a tissue specific fashion for example and we started thinking what if we could just do that by pre assembling a protein RNA complex much as we do in our biochemical lab but decorate this protein with ligands chemical ligands that would make it look like a car go for a receptor that's exposed on the surface of particular types of cells and so we we've been since to sort of bring us get us closer to thinking about how to do that we started working with Jennifer puck and Alex Marcin to immunologists at UCSF to think about how we could introduce these kinds of molecules into immune cells which have you know been very difficult to work with especially for genome editing because they're not easy to transform or transect and so the things that we've found is sort of summarizing some of our results with Alex marcin's lab is that we can detect editing of t-cells within a few hours or other kinds of cells as well so it can introduce these protein complexes right now we're doing it just by nucleo affections so we introduce them with a little electric shock to the cells but we want to in the future do it chemically or even through a receptor mediated pathway we find that the half-life of this RNA protein complex or RMP is about 24 hours so it minimizes off targeting and we really don't detect off target effects in these cells and finally we can actually Co deliver DNA templates for repair so we can really enhance the level at which we get recombination that introduces new information at the site of the repair which is a real advantage if you want to think about fixing a gene for example and so we've been doing this using initially using primary human t-cells and been able to do this for making changes in these cells to study the function of these particular genes but in the long term we really want to move towards towards therapeutic editing and thinking about how we can do this in of course things like I'm not up weeding stem cells and so I want to just in the last couple of minutes just cycle back to this question of ethics and I for me this really you know was something that came about somewhat slowly you know initially I was just very excited about all the science that was going on and in our lab and others with this technology but really you know the question kind of increasingly was coming up you know really how should we think about a technology like this and what should we do now that genomes can be edited relatively easily and for me it really came home when in early 2014 when a group published a paper in which they were able to edit the germ lines of and show that they could get genetically modified monkeys that not only had changes to their cells in their body but they could transmit those changes to their progeny so it you know and I started thinking well if you could do this monkeys mice rats and you know probably do it humans and you know when will someone try to do that and so that kind of motivated me to get a small group of scientists together we met in the Napa Valley in January of this year David Baltimore was there and Paul Berg both of whom had been involved in discussions about the ethics of molecular cloning back in the 70s and that remaining resulted in publication of this piece which was a sort of a you know a perspective in science magazine in which we proposed what we call a prudent path forward and we recommended that scientists not proceed with any clinical application of human germline editing in human embryos to give time for the community to get together and discuss this more more broadly and certainly globally and so as Doug mentioned this led we were very pleased that the National Academies took up this issue and last week in Washington we had a meeting that was sponsored by the National Academies of science in the u.s. the in China and the UK to discuss this and this I think has really kicked off a more global conversation about how to use this technology in ways that will be beneficial to human society and also respect a different group different societal norms and and ethical values so stay tuned because that will certainly be an ongoing discussion and I just want to also mention that I've been involved in catalyzing the formation of this group the innovative genomics initiative which is a UC Berkeley UCSF collaboration it's an academic endeavor that seeks to use genome editing for different kinds of applications especially in human health but we hope in the future in other systems and sponsoring research that is funded by both philanthropy donors as well as by companies so check us out on the web and I want to close by I don't know projector went out but but this is my most important slide because it's my thank you but ok so I'll have to see press here to show that on screen is that gonna work ah wonderful okay so this is very important because this is the the wonderful group of people involved in the science that I talked about and I mentioned many of them along the way today this was a picture taken at our lab 20 years sort of lab anniversary trip to my hometown in Hawaii last year and we don't do that every year I have to save my money for another 20 years to do it again and and I certainly want to thank collaborators I mentioned them along the way but I want to again highlight you know wonderful work that we had the chance to do with Emmanuelle and and her group especially Christian Linsky in her lab also Evan Agulhas my colleague at Berkeley for all of the e/m structural work that we're doing Eric green for the single molecule studies and and also Alex Marcin and Jennifer Puck at UCSF for the the more applied work and and then of course we couldn't do anything without money and again I want to really highlight you know we had some early support from the Gates Foundation we had support of course from I'm very grateful to the Howard Hughes Medical Institute but I also really want to give a special shout out you can't quite see the emblem here but for the National Science Foundation so NSF gave me a very small grant early on back before anybody really even had heard of crispers to hire the first student that came in and started doing curiosity driven research that led in obviously a totally unexpected direction so that's my plug for you know basic science funding I think it's really critical to pursue your passions in science and you never know where it'll take you thank you very much

Channel: David Geffen School of Medicine at UCLA

Views: 12,075

Keywords:

Id: eq6wpFREoso

Length: 59min 36sec (3576 seconds)

Published: Fri Dec 18 2015