Dr. Feng Zhang speaks at the Canada Gairdner Symposium: RNA and the New Genetics

so uh it's my pleasure to introduce our first Speaker this afternoon Dr Fang Jang um I'll keep the introduction for biographical things brief because you all have that in your book but he is a core member of the broad Institute of MIT and Harvard as well as an investigator at the McGovern Institute for brain research at MIT and the w WM C Career Development professor at MRT MIT and he is a joint appointment in the Departments of brain and cognitive sciences and biological engineering so the focus of his research is to understand the molecular Machinery of brain cells through development and application of Innovative Technologies and he's going to talk to us today about one of those innovated uh technologies that that he's been involved with so he's very interested in how the brain works in Neuroscience questions and has spent much time and effort uh and much productivity in developing uh Technologies to help address those questions and so those two technology uh that are uh so important in this uh well in all kinds of uh biology these days are optogenetics uh which is not the focus of today but of the genome editing uh mechanisms using crisper Cass and that's what we're going to hear from him today so he's developed uh crisper Cass systems to enable new cheaper and more effective ways to manipulate mamalian genomes and his Focus as I've said is on the brain and potential applications uh well beyond neuroscience and he been very very uh helpful to the community in making reagents available and in tips and in helping people uh with the use of these new and exciting and very powerful Technologies um so uh he's won he's recipient of many prizes and awards uh which I won't uh mention all of them he received his uh in chemistry and physics from Harvard College his PhD in chemistry from Stanford uh and he's now holds the positions that I mentioned at the beginning so the title of his talk is development of Christopher Cass 9 uh for genome editing thank well um good afternoon um it's it's really a great honor to be here and I want to thank uh John um for inviting me and and it's um uh really really really honored to be here so yesterday afternoon I had an opportunity to to meet with some of the students here at University of Toronto and I was really impressed with all the students and one of the things that they were very interested in is is where I grew up so I I actually had made some slides for a different talk uh previously so I just stuck in a couple of slides here so maybe I'll take about 60 seconds just to to show you a few things well I was born in in the 1980s in China and it was a time when China was just undergoing dramatic transformation and so this is a slide this is probably uh some of the things that you may see if you're walking around the streets in China um there's a strong emphasis on science and technology and I grew up in that culture so fortunately U my parents my mother immigrated to to the United States and so in 1993 um I was lucky enough to move uh from China uh all the way to to the US us and and there um I really um owe a lot of my education and training so so China and and the US are very different so here is what China looks like there's billions of people and this is what I saw when I landed in Iowa I exaggerated a little bit de MO is actually a fairly civilized place there there are fantastic school systems and so I went to a public school called Kalia middle school and this was really my first foray uh to molecular biology there was a Saturday in Rich program um and one of them was on molecular biology so I went and they showed us a documentary and this is a documentary so so from then on U my parents are computer programmers and and computer engineers and so growing up I was always interested how to program things and Jurassic Park told me that biology is potentially another programmable system and and it made a huge impression on me and also my mentor from that class remembered it so by the time I was a sophomore student in in in high school he approached me and said there's this fantastic opportunity at Iowa gene therapy center Research Institute and what about going there and trying to learn some molecular biology and so I I um I learned tremendous amount John Levy was my mentor very patient uh sat down with me every afternoon just to teach me uh various things he he took out sheet the papers and can draw all sorts of fascinating biological processes and many of those things I learned uh from then I still use today even when I uh show my graduate students how to do certain experiments so following that I went to Harvard and then had um two additional really fantastic mentors there's Don Wy a crystallographer and also sh J who I work with for three years to to learn how to do biological Imaging so it's during undergrad um both through seeing my friend suffering from psychiatric disease but also uh from me just wanting to learn understand more about my own mind that I started to become interested in neuroscience and and that's when I wanted to to focus my research on trying to understand brain processes and also try to help with neurological diseases so I went to Stanford and then worked with Carl daero um on developing a technology called AOG genetics but following graduation um I wanted to do something that's a little broader and so this is really the start of of what I wanted to tell you about today so here's a list of neurological and psychiatric diseases uh what is common about this disease is that we don't understand the molecular mechanism nor do we have um very good treatments for for most of them so fortunately the work of many many pioneers have revealed that um there are underlying and deeply rooted genetic and also likely epigenetic mechanisms for these neurological processes and so if we can figure out a way to systematically unravel these mechanisms then we may be able to develop a new squid of treatments to solve these diseases and so one of the venes that we can really capitalize on is the advancing human uh genome sequencing so here on the right is a plot from the National Institute of Health and it shows the number of genomewide Association studies that have been published over the past decade and you can see there's exponential growth Trend and each one of those data points each one of those papers uh represents at least one or if not many more hypotheses that need to be tested and indeed these genetic variations uh imply correlations but they don't imply causation for this disease and so one of the things that's important U for us to do is to be able to more rapidly and systematically go through and try to figure out what are the causal genetic mutations so one way to do this is through the generation of models and and how do we go about generating models well if you can generate a cell or animal model that carries the specific mutation that you want to study and then compare it with the wild type animal or cell and then you can do AB testing and try to figure out if this specific variant really plays the causal role in the disease State and we can do this um fairly well in mice um there's um Gene targeting developed by Mario keki and and others but to do this systematically a high throughput in any organism we want has been quite a bit of a challenge so one way to solve this challenge is through the the use of genome editing Technologies and so genome editing uh you uses a simp a tool called nucleases and the idea is is very simple and it was really um uh discovered by Jim Haber and also Maria Jason more than two decades ago what they found is that if there is a double strand break in a specific location in the genome then this double strand break can trigger uh DNA endogenous DNA damage repair processes there are two major Pathways that um we take advantage of one is called non homus enjoining and this is the eror prone process it's shown on the left side when the two pro broken pieces of DNA um are are are detected by the HJ machineries they will get rejoined together and and often times that will lead to the formation of a small insertion or deletion mutation we call this the indel mutation so if the indel is introduced in the coding region for a protein uh this can lead to a frame shift mutation and and uh allow us to knock out the gene function this works very well and uh we use this um in many other are now using this to generate knockout models the more precise method and even more powerful is the HDR repair process which um has already been uh taken advantage for generating a knocky knockout Mouse ESL and subsequently knocky knockout Mouse models um but here if we can use a nucleus to generate a s specific DNA double strand Break um this will allow us to increase the efficiency of HDR targeting by several orders of magnitude and when that happens we can provide a repair template and be able to switch in or remove uh specific sequences with much higher levels of efficiency and so in order to realize the potential of genome editing uh what needs to be developed are ways to be able to programmably introduce DNA double strand braks within a specific site in the genome and so to do this there have been many Technologies developed but they all hinge on the underlying principle uh which is having a DNA binding domain that is programmable and so here is one schematic of of one implementation of this technology whereby taking a DNA binding domain that you can easily alter The Binding specificity of you can consider this as a as a as a genome positioning system a GPS and you can use it to tether enatic domains uh nucleuses transcription modulator domains to a specific location in the genome and if you uh recruit a nucleus to to the Target site then you can stimulate uh genome editing so a few different DNA binding domains have been developed um the the one of the earlier ones is called Z finger proteins and and a more recent one it's called tails and we just heard Ben talk about um using Talons for doing genome editing but the the two technologies are somewhat cumbersome to use because it's quite difficult to program a protein to be able to recognize a new DNA sequence you have to either go through elaborate selection processes or to be able to uh construct together very repetitive sequences uh in a very uh systematic way and so I started with in fingers and then later on I worked on developing the the talent system but when I was working on it I thought this is not easy to use can we come up with something that's a a lot simpler and allow uh will allow us to be able to Target many more genomic regions uh with ease and so there was other Solutions so nature is much smarter than us and it turned out that over billions of years bacterial cells have already develop a system to be able to achieve DNA targeting in a programmable way with significant ease and this is the crisper system so crisper stands for cluster regularly interspace short palic repeat which is a mouthful but crisper is a lot of crisper to say and um and so basically it's an Adaptive immune system where uh when a virus infects the bacterial cell there are small fragments of DNA sequences that are incorporated into that bacterial cell from previous infections that can get activated and transcribed into RNA to then guide enzymes from the crisper system to destroy the invading DNA and so using RNA as a recognition um Motif to be able to recognize the invading DNA through base pairing uh these enzymes can more or less Target virtually any DNA sequence that you want to Target so this is a simple picture and we often take for granted that this is really discovered by a large body of scientists working in many different countries and in fact um this is just a a brief summary of some of the major contributions to the field was first discovered in 1987 serendipitously uh when scientists were looking into the genom thecoli uh and found these repetitive sequences and then in 2005 um a a major finding that um that outlined that the both hypothesis that crisper is an Adaptive RNA guided immune system and then over the subsequent years it was demonstrated to be indeed an effective system in 2007 and then multiple groups worked out the components and also the molecular mechanisms of the chrisper cast system and So based on all these early Pioneers work we harness the cine system by introducing the cine enzyme as well as working out two different ways of delivering the RNA one by either delivering the crisper RNA array which consists of these Target sequences along with the trace RNA which was discovered by emanu sharpen or by um by uh re-engineering a a single guide RNA that was initially reported biochemically by by sharen and DNA but then U we figure out that uh you have to have additional sequences which would make it work in human cells and so using this uh you can then uh achieve RNA guided DNA uh targeting and then be able to uh facilitate genome editing so our work on this started um actually was inspired by a paper that was published by Canadian researcher U his name is uh s muu so back in 2011 when I was just starting my lab um I had heard about chrisper through one of the meetings at the BR Institute and then when I went to look at chrisper I saw this paper that was published from his group in this paper he reported that um the cast n system in back then it was called cast 5 is the RNA guided DNA indon nucleus and so in this figure he showed that um the the the green highlighted uh sequence is is the spacer from the crisper system and that's a RNA sequence that when when complements with the target DNA either it's a plasmid or a Fage will facilitate a very precise targeted DNA double strand break and so that's where that that uh break in the in the um Sanger sequence Trace shows so I thought this is amazing um if you can use RNA to Target DNA and to achieve size specific DNA cleavage this will be a lot easier to use than either Zing fingers or Talon systems and so there were other there were many scientists working in the chrisper field and so just a few months later emanel sharpener working her La at Sweden at the time uh reported that in addition to the the guide RNA and also the cine enzyme there's a third component of RNA that's required which is the Tracer RNA and so Tracer rnaa complements uh with the criser RNA U by base paring in the repeat sequence uh of the crisper RNA and it and this dual RNA Guys cast n in terms of targeting uh uh DNA and and providing uh uh degradation function for in infectious viruses and So based on uh these early work we thought maybe we can try to harness these systems and try to see if they can work within a ukara cell to facilitate Human Genome editing and so at the time the field was um working on trying to understand the crisper system and some of the application were using this to be able to um improve industrial starter culture strings to be able to um uh make yogurt cultures that don't get spoiled because they will get infected by Fage viruses but we thought maybe we can use it uh to to add a genomes instead to to study genetic variations and and potentially uh to be able to advance gene therapy so I had to do a number of different things and early on I wrote a few grants but they unfortunately weren't funded but one of them was funded so this one was uh submitted back in January 2012 and basically we proposed to introduce the various components that Iman sharpen uh described in her paper uh namely cast 9 UM the crisper RNA the trace RNA um and also the bacteria Hol RNA 3 which is thought to be important uh for RNA guided uh Pro for processing of the crisal RNA and so basically uh without thought we can express these different components the cast enzyme the crisal RNA the trace RNA and RNA 3 to to see if they can reconstitute an active crisper complex within human cells and so we started to work on characterizing each one of these components and the first thing we did is to try to see if we Express cast n and also the RNA 3 within human cells and and more importantly to Target them into the miman nucleus which is something that bacterial cells don't have and so on the on the top um we we use nuclear localization signals to Target cast 9 to specific to the nucleus of human cell and we found that you have to have two you have have to really uh cine is a large protein and you really have to flank it on both ends with the NLS to get eff efficient nuclear targeting and also for rna3 you have to have uh targeting uh or attaching the NLS on the C term in order to get it in effectively so that's those are the two protein components um this the second um the RNA component we thought to see if um they can be expressing human cells and So based on the RNA seek data that Emanuel had published uh she showed that there are multiple Tracer RNA isoforms that are naturally expressed in bacterial cells and so this is uh on on on the top here um so so um we took each one of these isoforms and and Express in human cells and and what we found is that actually only one of the the isopor was able to be expressed inside of the human cell so on the on the bottom you can see Northern blot U where the the top the longer Red Bar indicates the the full uh long natural isoform but that one didn't Express and it's a shorter one that's also found in bacteria that is expressed within the human cell so then we thought um let's see if this system can uh Target Human Genome so we put in uh we tried two different cat9 molecules one from ESP pyogenes and another one from s thermophilus and then by altering the crisper array by replacing that that blue uh em uh that blue spacer with a piece of DNA sequence that correspond to the emx1 genomic sequence of a human cell we were able to try to see if we can retarget C9 to cleave this piece of endogenous Human Genome and what we found is is very interesting so on the right this is a gel we use the acid called surveyor to detect modification in a specific Locus in in the in the human genome and so the rightmost line shows introduction of all four pieces so there's cast 9 there's RNA 3 there's the crisper RNA and also there's the Tracer RNA and then the the second to last Lane uh shows three components without rna3 being provided and what we found is that um you actually didn't need to provide rna3 within a human cell to be able to achieve uh RNA guided DNA targeting and one nice thing about the crisper system is that um it's a naturally motop flexible system so these crisper arrays encode multiple spacers that can Target multiple sequences uh for example multiple viruses and we thought maybe we can use this to to have crisper cast N9 Target multiple sequences within the human genome and so what we did is we we introduce uh three things we put in cast n the trace RNA that can sa safely Express in human cell and also a single crisper array that carries both spacer uh two spacers one targeting uh Gene the Gene emx1 and the second spacer targeting the neuronal inter marker parment and what we found is that this system can work pretty well for targeting multiple genes using just a single uh piece of crisal RNA and so when you have both of these spacers you can detect robust modification in in both of these Gene targets that's the first two lanes uh in the gel and then if you only provide individual spacers then you don't see modification in in the second Locus suggesting that this indeed is a specific RNA guided genome editing uh uh effect so while we were working on this there were also other groups working on uh studying the C9 system as well and so uh v v sixes in Lithuania and also sharena now collaborating with Jennifer daa had published a very nice biochemical study uh showing that uh you can purify cast n and cleave DNA in Vil and one thing uh that they uh did is is they try to figure out what is the minimum but sufficient component for RNA guided DNA cleavage within within the human cell with within within a test tube and so what they showed is that um if you look at the CRNA and also the trace RNA the trace RNA can be substantially truncated into that just that very short fragment that's that's on the bottom hybridizing with with the the CRNA and so taking this information uh then they um show that you can actually link the two rnas together uh to form what they call a chimeric RNA and when they tested this in visual to Cle DNA in a test tube they found that U whereas the the last Lane is the chimea RNA where where you fuse the two pieces together the second to last Lane is the the full length Trace RNA so so that's not just the short blue fragment but the whole thing along with a CRNA that can Target the Target DNA sequence and so it so it very convincingly they demonstrated that chimeric RNA can work even more robustly than if you put the two rnas together um uh in in the cell so I thought this is great if you can simplify the the the rnas to just two components you can make the system even uh simpler so we thought to compare um the the Dual RNA system that we're working with with the the single chimic RNA and so this is targeting the emx1 locus using the Dual RNA so with the crisper array and and the full length Tracer RNA so two RNA pieces and this is what we found which we were very surprised by which is that the chimeric RNA with the components that work very robustly in visual didn't have activity within the human cell and so by comparing both of these RNA structures we thought maybe the secondary structures on the end of the TR RNA uh which are not essential in vitual might actually have important functions in vivil and so based on the hypothesis we started to uh lengthen the the the chimic RNA and what we found is that as soon as you started to include the first or both of the herpin then you can uh very dramatically rescue the activity and so now um based on the crystallization work that we had done in collaboration with osam Niki we know that these secondary structures are actually important for interaction of the guide RNA with with the cast 9 protein and so just to sum it up uh you can you can do genome editing um using two forms of of the RNA system you can either put in cast 9 with the Dual RNA system one benefit of that is that you can encode multiple spacers into the crisper RNA array to be able to Target multiple sequences at the same time or you can provide a single guide RNA uh with with the F length Trace RNA so since then we had adopted a a three PR strategy to to further develop the crisper cast system so the first is to um to further Advance crisper biology so that we can better understand the system and to be able to improve the specificity the efficacy and and also various other properties of the technology but at the same time we wanted to make sure the technology is is useful for a broad range of biological applications and so we try to work on developing a high through screening capabilities using the RNA guided IND nucleuses um rapid way to be able to generate mouth models or or even higher level animal models and also developing ways to be able to deliver this invivo to study Gene function in an efficient way and then the third is to to share the technology make sure that it's is open and widely accessible uh to really enable the community so in addition um to using C9 as a nucleus there are also other ways of using the system uh you can inactivate C9 uh to use it as a DNA RNA guided DNA binding domain to recruit effor domains to a specific spefic Locus in the genome you can turn genes on or or repress genes or you can put GP on it to be able to visualize things but another thing that's very exciting is is to U be able to um do genetic screens where um by uh synthesizing 60,000 or 990,000 uh guide uh sequences uh be able to very comprehensively uh perform loss of function or gain of function perturbation in the genome to identify new genes of in interest for a given biological process so we have been working on also uh trying to improve the the cast n system and one of the the primary uh concerns is the targeting specificity and so I'll show a little bit of unpublished data So based on the crystal structure of cast 9 uh what we um what we observed is that the in the inside the protein that interacts with the RNA and the DNA is highly positively charged and that um gave us uh the thought that maybe the way that cast Cleaves DNA is by using these positive charges to be able to stabilize the denat DNA so that it doesn't res it back up together and and allow the RNA uh to to invade and so if this stabilization is too strong then even if the RNA doesn't fully match it can probably still hold the DNA open and allow cast 9 to to make a cut and so without thought maybe if we weaken these interactions by weakening uh these uh positive charges so that the AR the DNA will more likely Z back together we can try to improve the specificity and so we generated um a number of um targeted mutations within this region and what we found is that the effect is quite dramatic and so on the bottom are some of the data from from this study where on the left side the the first set of bars show the wild type cast n protein and basically um if you look at mutations so so the two blue letters uh show mutations uh that are uh mismatches introduced along the length of the the guide RNA and what you see is that um mutations that are farther away from the three prime in are fairly well tolerated by the wild type cine protein but if you introduce these um substitutions that neutralize the positive charge within this groove U for the enhanced SP cast 1.0 and 1.1 you can see we more or less overcome uh these types of non-specific interactions for Cleavage and so this is U this is a subset of data and and what we are now working is trying to really understand what are the the best combinations of residues that we can use to to take this approach to to improve cast n and I think an approach like this will allow us to to Really significantly uh improve the specificity and precision of genome editing so for the rest of uh today's talk I just want to share with you some of the more uh most recent work that we have done on on on the Discovery and harnessing of new enzymes for doing genome editing uh nature is is is amazing and it contains many diverse organisms and we thought um there must be there might be other um powerful uh enzymes that we can harness to further advance and expand the genome editing toolbox and so the first story is on the development of uh shorter smaller more compact cast NES for invivo delivery and so so David Cox is a is a MD PhD student in my lab and uh and and he he wrote a very nice uh review describing potential therapeutic applications of genome editing and he summarized it as more or less just two um approaches one is xvivo delivery and one is invivo delivery so for diseases where the infected cells can be taken out of the body and repaired and then put back to the patient um we can do uh xvivo delivery so and there are many different methodologies available for introducing uh enzymes or or or DNA vectors into into cells uh in a peach Tre dish but that's only amenable to a small set of diseases mostly things that affect the blood system but for other uh diseases um we cannot take the organ out of the body we can um take the the brain out of a person fix it and put it back and for those U we may have to do invivo delivery and turns out that um as pyogenes castline which is the most commonly used one is is very large and it makes it challenging the package into a single anal associate virus vector and so we thought maybe we can make something that's more Compact and so we had taken two previous efforts uh to compact cast make it smaller one of them is to uh look for uh to to work with this other cast n from strep Theus thermop which is about 3.4 KB long but it turned out that the activity level of this particular cast n is not as robust as as pyogenes and the second approach we took was then to use structure to guide the the removal of access domains that may not be critical for cast function but for that we also found that uh we can get some success and reduce it by maybe 400 500 base pairs but that's not small enough and also it weakens the activity of the enzy and so thought well this is not good enough we really have to come up with a way that's more efficient um and and and maybe we can uh seek inspiration from nature again and so one thing we did is we we looked at all the cast nines that exist in in nature in sequence databases and then we plotted their length along along the the the x-axis and so what you see here is that cast NES fall into two length distributions there's the the smaller cluster that centers around a th amino acids and then there's the longer cluster that Cent is around 1350 amino acids and so ESP Pagen cast belongs to the longer cluster the the 1350 amino acids and we thought maybe we can harness the smaller cast N9 that would work well um in human cells from from the shorter cluster and so what we did is we we U did a philogenetic reconstruction of a shorter cluster and we picked six different short cast NES that kind of well represent this uh the diversity in this small small cluster so what we found is that um oops so what we found is that if you uh test the activity of these small enzymes in v um by expressing them in human cells and then licing the human cell to take the the the cell lysate all the small castes are functional within human uh cell within the human cell I say and and we can figure out what the Pam sequences are and we can see a pretty robust uh DNA uh cleavage but when we actually test the genome editing uh in human cells in living cells we found that only one of the six cines worked and that is a cine from um uh from sta caucus orius and that's the third lane that you see there so s orus had comparable activity as both the pyogenes and also the thermophilus cast n um but the other five didn't have uh the same level of activity and so that allowed us to then Package cast 9 along with the guide RNA into a single anal associate virus we we focus on aav because it's it's the only viral Vector that has been clinically proven uh and and is is is in a Jo called gbera in Europe and so we thought let's see if this can allow us to achieve efficient invivo genome editing and so what we did is we designed this guide RNA to Target pcsk9 which is a prominent drug Target for treating hypercholesteremia and so we thought let's put this in your mouse and see if we can knock out PCS K9 using sa C9 and then observe a phenotype and so this is the the data that we got um so on the the first plot on the left side shows um detection of indel formation in the pcsk9 in the liver of the mouse and what we found is that the effect is quite robust So within seven days of a Administration we observed over 40% indel formation within the liver tissue and then this modification efficiency more or less saturated and it persisted for the remaining time points that we took and if you take serum from the mouse to check both pcsk9 protein levels which is the middle plot or check the cholesteral level which is the the third plot uh what we see is that um within 7 Days of injection there's also a sharp drop both in the protein level as well as in the uh in the RNA uh both both in the protein level as well as in the the cholesterol level and so this showed that uh indeed we can build viral vectors that are Compact and can can mediate efficient inval delivery and and Achieve therapeutically relevant editing uh efficiency so so so that's to develop a smaller cast 9 but but there are other things that would be great to also address for example cast 9 generates double strand breaks and it's very good at knocking genes out um but um for inserting genes especially in postmitotic cells where the HDR process is not very active um we we really need to try to see if there are other possibilities and so we ask are there other enzymes out there other than cast 9 that can also provide genum uh editing functions maybe with expanded functionalities so if you look back to um the classification of the crisper system the the latest one that was published by Eugene Ken's lab um reclassify crisper system into two different classes there's class one which are the previously referred to type one and type three systems and they use multiple proteins to recognize the guide RNA and then also to provide the DNA or RNA targeting and destruction function class two is a simpler one where cast N9 belongs to uh it uses a single enzyme to recognize and Destroy uh invading uh DNA and we thought if there are multiple types of class one system there must be multiple types of class two systems as well and so turns out there are and and so working with Eugene kunan uh we examine one of these uh alternative uh class two systems and this is a protein called cpf1 so it turns out that in the bacteria francilla Nova there are two crisper Loi one that contains castai and another one that contains cpf1 and if you look at the Domain organization comparison between these two proteins you can see that cast 9 contains two nucleus domains there's the H&H domain and the r c domain RC is split into three different pieces that that come together when the protein is folded cpf1 contains a single non domain uh RC that's contiguous at a c term of the protein and so we thought let's see if uh this system is also active and if it's also a real crisper uh system so after characterizing the biological function we found that it indeed is a RNA guided DNA indon nucleus and there are three uh differences between that and cast n the first is that the target sequence um is the the guide RNA molecule is almost like the mirror image of the C 9 crisper RNA the um for the for the Target side um the Pam sequence is on on the left side and for C9 the Pam sequence is on the right side cpf1 uses at Rich pams and C9 uses uh GC Rich penss the second difference is that cpf1 is a single RNA guided indon nucleus it just uses a single short CRNA to provide the recognition whereas C9 is a dual RNA guided system uh you have to have this the crisper RNA as well as the the Tracer RNA and then the third is that cpf1 CLE DNA in a staggered fashion it generates a sticky end U very much like a restriction indon nucleus Hindi 3 bamh1 claw one so forth cast 9 generates a blunt in DNA cleavage and so this particular difference U may be advantageous for trying to acheve gen assertion in postmitotic cells because the nhj process is is very active in neurons and if we can provide repaired if we can provide donor sequences that carry compatible overhangs then we can potentially use nhj to insert sequences uh in in a postmo Cel so we're very excited and we thought to see can cp4 and mediate genome editing efficiently we tested Francis Sila cpf1 but we found that uh we were very disappointed it didn't show activity in a human cell but based on what we learned from cast 9 we thought well maybe FN cast FN cpf1 couldn't work but there may be other cpf ones that can work so we went ahead and we um um did a phog genetic reconstruction and we picked 15 additional cpf 1s that captured the entire diversity of cpf1 and then we tested them in in in human cells so these are uh the the 16 so FN plus the 15 additional ones you can see the sequences are actually quite diverse so when we're testing human cells um we found that even though all these cpf ons can work um in vual um so that's the gel the human cell lay gel that you see every single one cleave DNA robustly but when you tested in Vivo genome editing only two of the orthol worked one from Cena caucus and the second one from Lao spiria but nevertheless um this shows that um the this just shows that distinct bacterial species may have uh different environments that would render translation difficult and and it's important to find bacteria that have compatible intelling environments that would permit uh easy uh harnessing into human cells so so we also compared um the cpf1 uh efficiency with cast 9 and so we targeted four different loai and we found that uh they have comparable levels of genome editing efficiency so so this is cpf1 and cast 9 and there are now two different enzymes but are there additional enzymes we thought maybe we can carry out a search to see are there additional um crisper effectors that can also be useful for extending the genome editing toolbox and so working with Eugene kunan and also Constantine spov at scch we set up a computational algorithm to really mine through all the genomic sequences by looking for any unknown large Gene that's next to the crisper adaptation protein cast one and based on this um we found some new ones so here are the two known Ones cast n cpf1 and then there are three new ones that we just recently reported there's C2 C1 C2 C2 and C2 C3 C2 C1 and C2 C3 are are more similar in terms of domain organization at cpf1 but C2 C1 is a dual RNA guided nucleas whereas cpf1 is a single RNA guided nucleuses and C2 C2 is is the most uh interesting one because it has a very distinct organization so there's going to be uh many more to come so we're continuing along this three PR uh development strategy but what what we're most excited about is that we're now ready to apply this to study nervous system function and also mutations and diseases and so one of the things that we're working on is to be able to stud mutations that underly neurological and psychiatric disease and so one is is in visual modeling by using stem cells and putting in mutations to study whether or not they cause specific uh phenotypes within within the derived neurons and the second is to generate Mouse models of the same mutation and and make comparative studies to better understand the function of the gene within Inta circuits and also within the behavior of animal and so here are just two very quick examples so about 3 years ago it was reported that chd8 is a novel uh is a novel denovo mutation that's highly linked to autism spectrum disorder and so by using um genome editing to introduce this chda loss of function mutation into stem cells were able to derive neurons and then carry out electrophysiological characterization and so what you see on the bottom left the decrease Evol current response plot is a comparison of the firing threshold of these cells uh with uh with respond to input uh electrical current so what we found is that um the knout chd8 um cells have uh have a more immature like firing pattern and then we can also do the same in Vivo so we also build Mouse models that knock out chd8 and by doing the social interaction test that that Ben showed earlier um we found that there's also alter levels of social interaction Within These within these animals so these are just things that we're working on out that we're really excited about and that will hopefully help us better understand uh disease but in the long run what would be really exciting is to to Really refine uh the genome editing toolbox so that we can treat this long list of devastating genetic disorders and we working hard uh to to make it possible that sufficient specific and also to be able to deliver it uh to to the relevant tissue so this is the final uh most important slide I just only take 30 seconds um I want to First acknowledge all the students in post in my that uh are both passionate and also extremely creative in solving uh this problem but also our wonderful collaborators um Luciano who we worked with uh fairly early on on on the cast night system uh Osamu to to solve Crystal structures Eugene constin to uh look at various uh uh um uh new enzymes and and and many other collaborators on on applications and also uh many of my mentors who who played a critical role in uh in in helping me um uh learn how to do science and and to be able to um to work effectively as a scientist but also all the funding agencies that that supported us early on uh for these kinds of high-risk potentially no payoff uh type of uh uh project so thank you so much we will take questions okay thank you very much for this uh wonderful tour through this exciting technology we've got time for uh two questions we'll leave uh major questions for the uh uh Round Table after but if there's uh a couple of questions were Dr Zang is happy to you speaking next nobody oh Che uh yeah others probably can't though so let's get the microphone thank you um one question I had was I know uh in the review paper published by Dr Cox I believe was he mentioned um there's research going into potentially uh immune reactions because it is microbial derived um I don't know if you could comment on that a little bit yeah that's actually one of the um most important challenges facing clinical translation um these these cast 9 or cpf1 enzymes are microbial uh microbi derived and um and we we we um we haven't done enough characterization to understand the immune reactivity to the protein but for long-term applications uh that is certainly something that that need to be addressed maybe first um taking EP uh taking fragments of protein and figuring out what are the most endogenic epitopes and then trying to uh optimize them so that they don't evoke strong immune responses

Channel: CanadaGairdnerAwards

Views: 10,374

Keywords: Gairdner Foundation, Canada Gairdner Awards, CRISPR, RNA

Id: yT5JeLVUgdo

Length: 43min 57sec (2637 seconds)

Published: Tue Nov 24 2015