 Thank you, everybody, for coming back today for Ava's second talk. So again, this is part of the Steenbach lecture series, and these are held in memorial of Harry Steenbach, who was a professor here at the University of Wisconsin-Madison. Through his work on vitamin D research and patenting that intellectual property, he raised a fairly significant amount of money to help found the Wisconsin Alumni Research Foundation, which has provided the funds for this lecture series. So these funds include the Steenbach research lectures, as well as funding for Steenbach Library, endowed professorships, and startup funds for new professors. So we are thrilled to welcome back Ava Negales for her second talk today. I'm going to briefly reintroduce Ava, hitting a lot of the points that we talked about yesterday. She did her PhD at Keel University in the UK with Joan Bortus. She then did a postdoc at the Lawrence Berkeley National Laboratory with Kenneth Downing. And then she became an assistant professor at UC Berkeley, where she has remained since. She is currently a full professor of biochemistry, biophysics, and structural biology at UC Berkeley, and a senior faculty scientist in molecular biophysics and integrated bio-imaging at LDNL. She is a member of the National Academy of Sciences, a biophysical society fellow and a fellow of the American Society for Cell Biology. So we are thrilled today for Ava's talk, Regulatory Mechanisms of Polycomb Repressive Complex 2 in Gene Silencing. Ava. Thank you, Robert, and thank you again for allowing me to give these lectures. I'm honored and it's been two days and still more to come, a very stimulating talk, talks to faculty and students and hearing about all the wonderful science that is going on right here. So thank you very much for making me a participant. So yeah, today I'm going to tell you a second story within the theme in my lab that concerns gene expression regulation. In this case, telling you about the silencing complex, Polycomb Repressive Complex 2. So Polycomb is an essential complex to establish and maintain cell identity, so very critical during development, but also during the life of the adult organism where mis-expression, wrong expression levels or mutation can give rise to loss of that identity and actually contribute to cancer. So PRC2 is actually a major target as an anti-cancer therapeutic. The complex function by modifying chromatin, it trimethylates lysine 27 in histone H3 and this modification downstream ultimately leads to chromatin compaction and gene silencing. So the complex has four core subunits of which ECH2 is the one that has the methyl transverse activity and another one is ED, this is this accessor reader of that same modification and then allosterically activate ECH2 and it's thought that this is likely to contribute to the spreading of this modification to expensive regions of the genome. And there's also SUS12 and RBAP46 or 48, this one also comes by other names. In any case, these four core subunits are thought to never be alone in the cell. They interact with other factors, core factors to give rise to two different modalities, main modalities of PRC2 and the one that we study, sometimes referred to as PRC2.2 has as protein cofactors ABP2 and Jarrett2. So these proteins actually contribute to the regulation of the activity of the complex that thought to maybe also play roles in the localization in the genome and therefore contribute generally to the activity and place of activity of the complex. The other players, very important in the regulation have to do with histone modifications. I already mentioned that the tri-metallated lysine 27 in age 3 that PRC2 deposits is itself an activator of the complex, but the complex can read all the modifications, I'll tell you about a couple of them that also regulate either increase or decrease its activity. The joint effects of protein cofactors and histone modification can give rise to an activation of PRC2, which leads to more of this modification and compaction and silencing, or it can give rise to an inhibition of the complex, which then allow for more open chromatin in regions that would therefore be transcriptionally active. So of course, as I told you yesterday, we use cry-EM as a major methodology where the purified complexes are quickly frozen, thin and frozen into a thin layer in which they remain in a hydrated state. I have to tell you, when we started working on PRC2, PRC2 was too small to be studied by cry-EM. It's 300 kilodaltons right now, it's laughable, but at the time it was really borderline of what was considered to be doable. So when Claudia Tiferey was a postdoc in the lab, he now heads the structural biology department at Genentech here in the Bay Area. He actually started the studies in PRC2 and he overexpressed the four core subunits in insect cells and he found that the complex was really, really flexible and we consider it intractable for structural studies. But of course, he knew about these cofactors and what he did was to add one more protein and that was ABP2 and this had a significant effect in the stability of the complex. I'll show you why in a minute. He used negative stain to characterize it and he was able to obtain a very lovely structure at 20 angstrom resolution. Now, if you look at the schematic of the structural domain in the difference subunits, you will see that both ED and RBAP48 are WD40 domains that at 20 angstrom resolution look like donuts and there were two of those in the structure. So that meant that out of that structure that he had produced, those two correspond to those two, likely to correspond to those two proteins but which one was which and what was the meaning of the rest of the complex at this resolution, what could we tell? But Claudio was very resourceful. He had a hydrologist expression system so he could play with it and this is what he decided to do. He decided to systematically place one at a time a label like either an MVP on the N or C terminals or a GFP in places between domains or between secondary structure elements within domains, a GFP in order to then compare it with a Y type, see the position of this extra density and therefore get an idea of where the difference of units were located. So this is the idea. We don't do this anymore but it was really amazing what he did at the time. So this is just two dimensional projection. If you concentrate on this one, for example, this is the Y type, this is the equivalent projection, the same direction of view but with a GFP that has been placed between the first and the second blade of the beta propeller of ED. And you see this extra density, also you can see it here, you can see it in the difference map that allows him to actually position that subunit and that particular place with respect to the three dimensional structure. He did this very systematically. In this case, he moved through the blades. This, for example, is a GFP between the last and one to last and it is in the same region but it's in exactly a different position. We could even place, do a docking of the beta propeller, orienting it angularly because he was able to place the GFP between different blades. I thought it was truly amazing. So doing this systematically, he was able to paint the different regions of the complex with the different domains. And at the point, there was only structures, I believe, for one of the WD-40 domains, for the others, the SUM domains, the SET domain, there were models that we could use for homology and just they were placed in the structure just as a placeholder. But basically, this gave us an architecture of all the different subunits wearing the complex that I'm going so much about because it turned out to be spot on. So it tells you when you do the right experiments that you're resourceful and tenacious, what you can do even when the technology is still limited. By the way, I don't want to forget, we corroborated this structure, collaborating with Rudy Eversall, doing cross-linking mass spectroscopy to make sure that what we were doing actually makes sense by something that was a completely orthogonal method. This is something that we've done for many other structures for TF2D, for the risk chromatin remodeling complex or for even higher resolution structures of PRCT. Yeah, very beautifully complementary. When your resolution is not so, so very high that you want to use something else unrelated to corroborate it. OK, so since we published that structure in 2012, not a lot happened for a while until at the end of 2015, beginning of 2016, two fantastic crystallographic studies came first on a fungi and then on the human complex that included what we have identified at the top lobe of the complex, which turned out to be the minimal unit for biochemical activity, both methyl transverse but also allosteric regulation. And that includes all of ECH2, EED and part of SUS12, so sorry. So that element and just with these two full subunits, the one with the catalytic site, the one with the allosteric site and the very C-terminal VEPS domain of SUS12 form a unit, that functional unit that was crystallized. And this is the structure. And one of the things that was very interesting is that they were able to obtain the structure bound to a histone peptide that had been trimethylated in lysine 27, bound to the allosteric site of the WD-40 domain of EED. And they saw that this activating peptide resulted in the stabilization of a segment that is invisible otherwise, but that falls as an alpha helix on top of the peptide and the EED that they call the stimulatory response motif or SRM. And this in turn interacts with one of the helix in the set domain and give rise to an optimally structured active site to receive a peptide and modify it in the active site. So this was the structured information that we have at the time. And then Vignesh Kasinath, who I realized yesterday talking to people was here not too long ago presenting a big part of what I'm going to be telling you today. So today you should really, really end up learning a lot about PRC2 by hearing the same things a couple of times. So Vignesh said he wanted to get the full complex and the full complex, meaning the full PRC2.2 with the two activators, ABP2 and JAR2. So again, he used an insect over expression system to produce these proteins. And he made progress, but he was for a long time stuck in the limbo of about four and a half months from where the regions that were novel were still not traceable. And what ultimately did it was the use of the Volta faceplate. This is one of the very few cases where the Volta faceplate has been used for another initial structure determination of something. Another one was our study on human TF2H. And there's not very many more because the Volta faceplate is very hard to use. And when it comes to really small things, it turns out that it's better to just use 200 kV scopes and get done with it. In any case, one of the things that I also want to... So this is how the data looks like. The contrast is unbelievable with the faceplate when it does work. The sample had to be crosslinked and deposited on a carbon surface because this complex otherwise breaks apart. So it's one of these complexes that really gets damaged at the air-water interface, which is an infinite hydrophobic sink. And it can compete for hydrophobic interactions in complexes or even lead to unfolding of proteins. So we have to both crosslink it and we use BS3, the same crosslink that we use for the mass spec studies in very, very small amounts of very likely crosslink. And a carbon support that again allows for the adherence of the complex to that surface so that it's less likely that it diffuses away and contacts the air-water interface. In any case, using that, using focus classification of the type that I told you about yesterday, he was able to get a structure about 3.5 angstrom resolution and this was enough to not only be able to use the initial structures from crystallography and then adjust them, but also for the top load, also to generate the structure of the bottom load that was unknown until then. Now, in this structure, he did not add any histone peptide, but it's a structure that is activated and it's activated because it turns out that Jarrett-Tum discofactor is itself a substrate for ECH2. And it gets modified and the tri-methylated peptide, that part of the protein, binds to the allosteric site in EED and activates the complex. OK, so what you see here is an activated peptide, it's not added, it's just there from the prep. These subunits are co-expressed in insect cells where all the chemistry is there for the methyl transverse activity to occur and when we purify it, Jarrett-Tum is methylated. And he has bound to the complex. OK, so actually, Vignette didn't see one structure, he actually saw two different structures that coexisted in our biochemical preparation. And these two estates differ on the position of this domain, which is the sun domain that comes after the sun binding domain, this helix, this SBD helix, that can be either straight, or bent, and when it's bent, the sun domain taxed in and when it's not, it's sticking out. So this is an orthogonal view of that. To tell you also that here you see the sun domain down with this helix bent. What you see here in magenta is the stimulating peptide, that in this case comes from Jarrett-Tum. What you see in yellow here is the SRM that has been stabilized and has folded. And I'll tell you in a second, but the active site is in an active conformation. So these are just details of the peptide and the trimethylated license surrounded by the classical hydrophobic cage being recognized by this hydrophobic cage. But in addition, here's this structure, this extended active conformation with this helix straight, the sand one in the breeze. And it does have a stimulating peptide, but we don't see the SRM. So the SRM is not folded over the peptide. However, the active site is also in an active conformation. In any case, this is the same region where you can again see the activating peptide. So this is telling us that there's an equilibrium between an SRM that is folded and unfolded. And somehow under our conditions, it seems to correlate with having this region up or down. Okay, so bent SBD, presence of SRM, straight SBD, no SRM visible. And just so that, you know, we can, this was just a control where we express a Jarrett 2 that is shorter. So the license that is methylated is 116. So if we generate a construct that is shorter, then it doesn't have that residue. And it is not that there's nothing occupying the allosteric site. And concomitantly, we see nothing occupying the active site, which is not in the right conformation, which otherwise, for the other two active, we see engage also with substrate peptide. Okay, so I just want to concentrate on one more feature of this structure. And this was the fact that we saw a part of ABP2. ABP2 is predicted to have a couple of small folding domains and all of long linkers. And this is the region that we saw binding to the complex. And I want you to see how it binds right up the interface between the top and bottom loop, which is the other ones that are very flexible in its absence, but in its present. It's like a staple that keeps the two, the two halves closer, you know, tighter together. By the way, I'm not going to show you this, but the interface between the top and bottom loop is really dominated by a hydrophobic patch. And we believe that this is the some breathing still in the structure and this hydrophobic patch tend to be competed away by the air-water interface that unravels Zeus 12 and then RBAP48 falls off. And this is the way the complex breaks. And I'm going to show you that broken complex in a second for a reason that would be that I would explain. In any case, the surprise, we expected that ABP2 were going to be here because it's stabilizing it is the most likely way you serve as a glue yourself to the regions that are mobile. But something that came as a surprise is that there's a part of ABP2 that we see very well and that binds to the WD-40 domain of the RBAP48. And this domain had been shown before biochemically and through a crystal structure to bind unmethylating H3 by the by lysine K4. I'm going to be telling you about K4 later on in an unmodified state. This WD-40 domain binds it with reasonable affinity. And what we see is that ABP2 can mimic the binding on that site. So interestingly, both cofactors are mimicking histones. Jarrett 2 mimics a K27 trimethylated histone and activates the complex binding to the allosteric site. Well, ABP2 binds to a site that normally recognizes unmethylated K4. The reason for it, we actually don't know what the function it is, but it's just an amazing coincidence that both of them are able to play these roles. So once we have the structure of the complex, the obvious thing was can we see it bound to this template, which is nucleosomes. And in particular, the first study that we did was not for a mononucleosome, it was for a dinucleosome because we wanted to get hopefully some information about how spreading may occur in a structural context. So this is the idea. This was the work of Simon Popsen when he was a postdoc in my lab. Now he has his own at the University of Cologne in Germany. The idea is let's work with a dinucleosome, where one of the nucleosomes has already been modified. It's already trimethylated on K27. So it can serve as an allosteric activator of the complex. And then the one next door will be unmethylated and could serve as a substrate. So the idea is that part of the spreading mechanism could involve PRC2 engaging both simultaneously and the activation of one nucleosome will therefore promote the methylation of the one next to it. So Simon generated these two types of nucleosomes. Then he ligated them to generate the dinucleosome. And then he first tested whether PRC2 was binding to the dinucleosome in the way that we expected. And this is the result. I hope you can see very clearly comparing the PRC2 by itself and the PRC2 with the dinucleosome that we can see both two nucleosomes engage simultaneously with one complex. So this was great. So it was the time to move to crime. And here came the problem. The problem was that this complex could not be cross linked. If we cross linked the complex, the complex comes apart and we see that both by negative state or by gel shift. So somehow the cross linking is incompatible. We understand now why the contacts are happening mostly between lysine and nucleosomal DNA. And those lysines were being titrated away by the cross linker. And this also told us because this is a problem even after we've formed the complex and we can see it. It just tells us that these contacts, there has to be a certain off rate and these things are coming on and off that the lysines can be competed away by the cross linker that binds to during the office date and then do not allow for any further rebinding. So he also could not use a carbon support because when he did, the geometry of the complex with the two nucleosomes was such that he only got one view. So he had to buy the bullet and do cry-EM without the carbon and without the cross linker where we knew the complex was going to be affected and this is what we could do at the time. So we did it and we could see the two nucleosomes and the top half of the complex which we know from negative estates is the part that binds the nucleosomes anyway but the bottom half was disappearing and that's what I was telling you before. He has still went ahead. He obtained a beautiful structure about seven, eight ounce room resolution where he could take the structures that we have for the top lobe and the Luger structure for the nucleosome model the linker DNA and get a structure of how the PRC2 contacted the nucleosomes. There was certain flexibility here. That's why the resolution is not very good. These are just two of the most extreme classes that are superimposed on PRC2 and show certain motion in what is, we know I'll tell you in a second is the substrate nucleosome and really quite dramatic motion in the modified nucleosome that serves as an allosteric activator. By the way, to start with, when I saw this structure, it was beautiful but we thought, is this a coincidence that we use the right, how lucky were we choosing the length of the linker DNA that these two things can engage with each other? So we changed the linker length to 30 to 40. The complex is still binds. Whether it binds with exactly the same affinity probably is slightly different but is able to do it. And for that he relies on the fact that there's two H3s per nucleosome so he can flip the nucleosome like in this case if flipped with respect to one another if flipped this nucleosome is engaging a different H3 but also with this flexibility that exists that gives you a little also leeway in the angle and therefore the length of the linker DNA. So how are these two nucleosomes being engaged? So first of all, let me tell you about the tails because that's how we know which one is which and they're doing what we're expecting to do. This nucleosome is the modified nucleosome. We can see the modified tail bound to the allosteric site. By the way, this is a five subunit component. There's no Jarrett two, ABP two is there but Jarrett two is not there. In this case, Jarrett two is not the one that is activating the complex. It is the modified nucleosome that is activating the complex, okay? And on the other hand, we see the nucleosome and we see the tail making it all the way to the active site in the set domain. But most of the binding energy doesn't come from the tails. It comes from interacting with the nucleosomal DNA. So this is how the nucleosome and I'm just showing again the range of motion at which it can be engaged is pivots with respect to the contact point. The modified nucleosome interacts with the ED and with this SPD, this long helix that then leads to the San Juan domain. Remember this, this is a straight helix. This helix in a straight form engages this nucleosome. It would not have the same contact if it was bent, okay? On the other hand, it was in the previous slide, the SRM, I wanna show you. The SRM is perfectly stabilized. So this is a mixture in the presence of a nucleosome. This is a mixture of the state that we saw before, one with a straight SPD and a San domain that is up and detached, but that has a stabilized SRM. Okay, sorry. And then, so that's the contact of the allosteric nucleosome, the one that activates. This is the substrate nucleosome that is proximal to the set domain and where the DNA is being contacted by the CXC domain, proximal to the set domain within ECH2. And in both cases, there are very nicely defined positively chart regions that interact with the two gyres of the nucleosome in the case of the allosteric pre-modified nucleosome. And with the two, you can follow nicely the two strands of the one gyre of DNA that interacts with next to the active site. All right, so this was great. It was a medium resolution structure that we could interpret with previous either crystal structures over on, but there was one element that was unexplained. So if you turn this around, the way I did, there is an extra density that sits on a really prime position between the PRC2, the ECH2, the nucleosomal DNA and the histone tail in its way to the active site. And this density shown a two different threshold, one that shows the density is very strong. The second threshold in white is so that you can see more or less the path of the peptide of the histone tail. So this is, it seems like a very strategic position is even more so because what you see there with that asterisk is K36. This is another lysine that gets primethylated by a different complex and it's inhibitory have been shown to be inhibitory for PRC2 activity. So what is that? I'll get to that in a second. All right, so, but I want to go back to this issue of the histone modifications. So there's modifications that are associated with active transcription that are inhibitory to PRC2, that's trimethylation of K4 and trimethylation of K36. This have been shown in the context of the core complex, which is by the way, very poor, have very poor and thematic activity to be inhibitory. On the other hand, there is the H3K27 trimethylation, the self modification that is actually allosterically activates the complex and there is another modification, which is the mono ubiquitination of lysine 119 in chisoon H2A. This is a modification that is introduced by the other major polycom complex, polycom repressed complex one. And once this modification is introduced, PRC2 has been proposed to be recruited by that modification to modify sites in the genome, then deposit the modification that I've been telling you about and this is itself recognized by a slightly different form of PRC1 that then gets further recruited to the site and it's thought that that PRC1 has the capacity to itself contribute to the condensation of chromatin. So we decided we wanted to look at these mononucleosomes bound to these mono ubiquitin in the context of the six subunit PRC2, but we wanted to be able to do better that we have done in the past, where the lack of crosslinking really got rid of the bottom lobe of PRC2. So this is what we did. So Simon and Vignesh got together and were successful in implementing a methodology that had been pioneered by Bob Glaser at UC Berkeley and tested with ribosomes by Jamie Kate at UC Berkeley. And the idea here is to use a two-d crystal a two-d crystal estreptabline monolayer to bind your complex. So this is the concept. The concept is that you have biotinylated lipids and a lipid monolayer where they are kept fluid. You add estreptabline, the estreptabline binds and the fluidity of the lipid allows it to move around and crystallize and grow a two-dimensional crystal in that monolayer. And then all you have to do is to biotinylate your sample. So this can be done in the default way is that you have your sample PRC2 and you in vitro biotinylated at very, very low level so that you typically get zero or one, maybe two license modified per complex. And those will be random on the surface. Then you bind into the monolayer and that will do a number of things. It will concentrate your sample. It will provide random orientation. And as I will show you for PRC2 by being able to then wash the grid so that the only thing that remains in there is the complex that is attached to the monolayer. Then when you blot and you get to a thin layer the complex does not have the capacity to diffuse touch the air-water interface and get damage. So it also is preserved against the deleterious effects of the air-water interface. So in our case so far, although we are thinking of trying the other way around, what we did because we had a substrate was to biotinylate one end of the DNA in our mononucleosome construct. So that binds to the monolayer we can and then we can add the PRC2 which binds to the nucleosome and then we can do our studies. Now you could do this without forming a two-dimensional crystal but the crystal has a benefit. So the septatin contributes to the noise if you want of the image. When you have a lattice which I don't know if you will be able to see with the resolution in your screen but there's a square lattice in here. If you fully transform the image you can get Brack reflection diffraction spots and you can mask those out in Fourier space which basically just remove filters out the septatin contribution. Okay, so that means that we have one additional step to do before we do any other image processing but after that we are left without any contribution from the septatin. And this worked very well. And this is the structure that Vignesh obtained bound to this mono ubiquitinated nucleosome and I'm gonna go into some details but this structure not only have preserved the bottom half it was actually improved in many respects with respect to what we have seen before in addition to give us the information on how this particular nucleosome was being recognized. So some of the features that I will mention some of them very quickly. So Vignesh was able to extend the part of ABP2 that was modeled to include a region an extension that includes the have a highly positively charged KRRR motif that Danny Reimber had identified as being involved in DNA binding and in fact it binds to the link DNA extending from the nucleosome and it's a stabilizing app that we can see it. But then so that's what I just told you the other one is this mysterious density that is present when the complex is bound to a substrate nucleosome but was not present in our cryome structures not present in the crystal structures than before in the absence of nucleosome. And we refer to these helix as the bridge helix because it bridges across from the set domain the DNA, the nucleosomal DNA and the histone tail itself. So this is a helix that the form is unstructured in an invisible in the absence of nucleosomes but it falls and contributes to the interaction with the nucleosome in its present. And there are more interesting things about this bridge helix and is that it includes one end, two license in ECH2 that are automatically methylated. So this automation has been shown to contribute to the activation of the complex. The methylation occurs in Cs, the activation occurs in trans. So that it occurs in Cs. I think the structure explains this end of the helix is very close to the active site and when the helix is unrubble it will perfectly reach the active site and be methylated. So the methylation occurs previous to binding to the nucleosome but then in fact, by the way the mass spectroscopy shows that our sample is automated in those license. So they're right here. How they activate in trans should be next chapter. Maybe I can come back in real person and tell you about that next time. Okay, and then one thing that we finally see really very clearly before we saw not so clearly is the path of the histone tail all the way from the core of the nucleosome into the active site all the way to the license 27 that is into the active site and a little bit beyond. So all of these is very stable in the complex. There's a path of residues that basically channel this very extended form of the tail right into the active site. And if you, very interestingly this is pretty important because if you map that is a richness of an enrichment of mutations that I found in cancer right at the residues that directly interact with the tail or that interacts in the second layer if you want on that helix that I tell you that it stabilizes upon interaction with the nucleosome and the H3 histone tail. Very cool. Now, but what is special about this particular sample is that it was mono ubiquitinated. How is the mono ubiquitination recognized? And the recognition is done by the two cofactors both Jarrett II and ABP II. So in here, the purple, so ABP II we see in pieces is a very extended protein with very long uninstructed region. So we see the peptide that is activating that I told you about and we see another that interacts with ABP II right at the neck between the top and bottom but now we see this region which corresponds to very close to the end terminus of the protein that is involved in binding the nucleosome the ubiquitinated nucleosome. So it had just a little before our structure came out had already been identified through biochemical analysis that there was ubiquitin interaction motif in this region. We see that it sits sandwich between the ubiquitin and the nucleosomal DNA but we see that the interaction is extended. There is a second helix, I'm gonna turn it rotated that extend from that and that actually binds over the infamous acidic patch on the histone core. So that this extended region is the one that is involved in binding to the ubiquitin and also binding to the histone core. Remember that otherwise, all the other interactions that we have seen before involve binding to the DNA and the histone tail but not to the histone core itself. There's another interaction that happens on the other surface but for the other H2 H2H histone or that second ubiquitin and that involves two out of the three zinc fingers that are present at the end terminal end of ABP2. So again, they bind wedging between the ubiquitin and the histone surface but in this case not the acidic patch but one that has a mixture of positive and negative charges. So those are the two interactions that contribute to the recognition of the ubiquitinated nucleosome, a mark that is inserted by the polychrome repressive one. So just a couple of controls that we did at low resolution but just to check that things make sense. One thing is that we did not see interaction with the histone cores in the absence of ubiquitin although there are regions that are binding to these patches on the histone core. So if we remove the ubiquitin and we look really, we don't see density for the cofactors on the histone core. So you need that the avidity that is provided by the presence of ubiquitin to zip those extra regions and bind to the nucleosome. With that ubiquitin those interactions do not exist. This is another a different type of control is on control in which we remove the end terminal region in Jarrett 2 that binds to the ubiquitin. So everything else is in there, the activating region is there and but you don't see the Jarrett 2 but also you don't see the density for the ubiquitin. The ubiquitin is the most flexible part in the structure without the Jarrett 2 is so flexible that we don't see it. The one on the other surface that interacts with AVP2 that one we see, but if we remove the contribution of Jarrett 2 that particular ubiquitin is no longer visible due to mobility. So those are a couple of controls. Okay, so now that we have these structures and that the recognition of ubiquitin is done by the cofactors we thought we will do some activity access because they had been contradictory information about whether ubiquitin, that mono ubiquitination was activating or repressing the complex but just like for the truly repressive marks the K4 and K36 trimethylation those experiments have been done with core PRC2. So when he said it, we will test activity for different complexes and with different modifications. So this is what I'm gonna show you now. When you see in different colors are the type of substrate it could be just the white type if you want a modified nucleosome, then the nucleosome that is trimethylated at K4 the nucleosome that is trimethylated at K36 and in green and then in this kind of magenta is the mono ubiquitinated H2A. Okay, so you can see that if you use the PRC2 core first of all is a very poor substrate no matter what but there is a statistically significant decline in activity in the presence of these two modifications that places of the ubiquitin doesn't do anything. If you include ABP2, the activity increases a bit for the unmodified complex but remains about the same. So if you want a more significant difference. Now this is what happens once you add Jarret2. Remember Jarret2 gets automatically methylated and activates the complex. So now this is a really active complex on the unmodified nucleosome. It has a lot of activity and there is a significant reduction in the activity for these two modifications that are identified with regions of active transcription in the genome. But I want you to see that this is far from zero. This is still pretty active. Then if you add ABP2 but without the zinc fingers that are involved in recognizing the second ubiquitin, you get this is somehow similar. I just want you to see that when you have full ABP2 and Jarret2, then you see this additional statistically significant increase in the activity of the ubiquitinated nucleosome. That maybe is simply explained by just a retention and the binding and the stabilization of that engagement by binding that extra binding to the ubiquity. But in any case, this tells us and this, I think this is important that these modifications, these two modifications that K4 and K36 are very similar to the previous one and K36 are indeed inhibiting but they're not eliminating the activity of PRC2. So we thought it was interesting to look into what these modifications could do and this is just front and back view of our structure with the mono ubiquitinated nucleosome but looking at that particular lysine. So the unmodified K36 in H3 is actually interacting in the complex where it's been kind of channel, okay? It's interacted both with the protein with ECH2 and with the DNA, the nucleosomal DNA, okay? It seems here that there's no space for trimethylation to occur, that it will be a steric hindrance but there really is nothing behind. So these sidechains could take a different conformer and that is in that case a space for that modification, maybe. What effect it will have confirmationally, I don't know. It would also lose these two interactions. So I think this is, it would be interesting to look at the actual structure with that modification and see what happened. So I just want to tell you that what we have started to do, we haven't got to the situation where we have high resolution yet, is to look at the trimethylated K4 nucleosomes, okay? So this, again, these were nucleosomes in which PRC do have half the activity that it has in the unmodified nucleosome and this is what we see. We see that the complex is partitioning between two states which are approximately 50-50, two states. In one state, the H3 tail is engaged and it makes it all the way to the active side so it looks extremely similar to what we see for the other structures but in the other half there is no density for the histone tail. The nucleosome is engaged exactly the same way. The bridge helix is there, there's density for it. There this is just docking the R structures from the mono ubiquitinated into these lower resolution structures. So there is density, there is an engagement of the nucleosome through interaction with the nucleosomal DNA but the tail is not visible. The tail is somewhere else in half of the complexes. So tail is engaged, there is something else that is happening. There are conformational changes that even at this lower resolution we can see. The sun domain and the region around the allosteric site there somehow change. And right now although we cannot tell, remember that this around here is about residue 24. So there's a bunch of residues that are there flopping around until you get to K4. And the trimethylation is somehow being red and that leads to some of the complexes not making into, not threading the tail. So I just think that there are the trimethylated K4 is being recognized somewhere else and is being sequestered away. But somehow the complex is still capable of competing it away and engage and have reduced activity but do it. So we really want to push the resolution here so that we can really understand how the trimethylated K4 is recognized somewhere else in the complex and sequester form from the active site. So this is what I wanted to tell you. I just want to thank everybody involved in PhD studies. So Vichnes now has started his own lab at Boulder while he was here. He worked with two very talented undergraduates Jennifer who is now doing her PhD at MIT and Curtis who's still in the lab but very soon is going to leave to do his PhD at Johns Hopkins. The whole story about working with PRC2 was really initiated by Claudio and Simon was the brave soul that decided to look at a chromatin template and actually did the work with not one but two nucleosomes in that initial study. And Paul is another postdoc now working in other aspects of PRC2 regulation that hopefully will mature and I will have the chance to tell you about soon after and I did not include it but I have a new graduate student Trinity Cookies that has also joined the PRC2 team very, very recently, officially in a couple of weeks. So we hope that there's many more concerning the regulation of this complex that we will be able to tell you next time. Thank you very much and hopefully today you have questions. Thank you for that real problem. That was amazing. I would probably look at your 2D class averages of histones all day long. So I have several questions. So part of my first question relates to this domain in SANT 1 that seems to be moving up and down that you originally found kind of correlated with the SRM but it didn't seem to correlate the SRM presence didn't seem to correlate with activity of the complex and you see the domains and SRMs in different confirmations. Yeah, so in when we work without an allosteric nucleosome we the most occupied by the state is one in which the SPD is bent and that SANT 1 domain is stacked in you can see right here in this particular structure. Once we have what we saw in the April which had been activated by Jarrett 2 is that this was partitioned between two states and that the SRM could be there or not. So it may be that whether you activate with the H3 or you activate with the H3 in the context of a nucleosome or you activate with Jarrett 3 those three gives you different states but you know I think what we see in the in the case where we don't have nucleosomes in Jarrett 2 is we can see that that capability is part of the conformational landscape that can be if you want utilize to regulate activity or maybe regulate even the interaction of PRC2 with other things that we haven't started looking into like RNA for example. So that change occurs it is important in the context of an allosteric nucleosome to contribute to the binding of that nucleosome. It doesn't seem to be of any much relevance in the context of activation via Jarrett 2 and that SRM this yellow thing that I'm pointing out right now that we thought was absolutely essential to have an active center in ECH2 that works is not the case because in the structure we see something that is engaged with a tail I mean we all activity assays whatever activity we see in cases where we have a mixture is very hard to distinguish one versus the other we will have to do really go and dissect it using mutations and things like that where we can make one conformation disappear and see what effect it has in activity because all the activity could be coming from one of the sides but just based on the structure even with that SRM in place we saw an active side that was fully engaged with peptide so you know I think it's very cool that there are these elements that can exist in different conformations and that can be relevant or not for activity in different contexts and that allows you layers to work together I think our next question comes from David Brow is binding a PRC2 mutually exclusive with H1 no no we've done that we haven't published it yet no it's not so you don't have a structure for it I take it on PRC2 bound to a nucleosome yeah and does that make sense with the structure you have without H1 I mean can you figure out how it fits in yeah we can see where the H1 goes I mean we have this nucleosome with two linkers and how it sits interacting with the linkers DNA and there's no hindrance or anything that is stopping the PRC2 for engaging the nucleosome I don't know if that's exactly what you're asking but yeah so yeah I don't have I don't have more details because I don't I don't have a we don't have a high resolution structure and we actually looking at this in the context of even larger arrangement that's consistent with the biochemistry I take it that it's already known that PRC2 can act on 30 nanometer filament or whatever um you know I think I would you know we've looked into modeling the arrangement of the nucleosomes in the 30 nanometer fiber to see whether this aesthetic hindrance due to that arrangement and right now I don't remember but I don't I think I would remember better if we have thought some big problem and I just I just don't but I'm sorry I should remember better and maybe we should we should rethink more now that we're in the context of that of that H1 thanks Bob you had a question yeah hi it was Bob Linnick it was a beautiful talk and I have trouble keeping up with all of the modifications and what they're doing and activating so I was wanting to ask a question to see if I can you can help me put this together at a somewhat higher level if I understood what you said in the introduction the PRC2 depending on what it's interacting with is capable of either spreading repressive or activating the effects on the nucleosomes sorry sorry no PRC2 if it's active will leave to compaction if it's inactivated it doesn't do anything so it doesn't lead to these modifications and therefore chromatin could be expanded you know there are there are sites in the genome that get repressed and they get kind of repressed forever that have some interesting mixtures of modification that can be switched and I think you know in some of these places you have PRC2 but it's not but there's no contents there's no the chromatin has not been condensed or you don't even have the K36 trimethylation but you can IP by chip you can see that there's some PRC2 there in the presence of this so there's something more complex that the super simplified kind of things that we've seen right now but that where we do one modification at a time and we are just seeing how well the complex is engaged and how well it is acting on a simple template but when it comes to being fully active inhibited or half inhibited what does it mean in the context of different sites with different epigenetic modifications it's just a little bit beyond what we are able to do it's another level of complexity so this contributes to at least saying we have to be aware these modifications which are they correlate with regions of open chromatin and active transcription they are not fully repressing PRC2 at least by themselves so it has to be that if PRC2 is kept repressed there has to be other players right it can be that the Jarrett 2 is kicked out in which case that complex is less active and then it gets further repressed so this combinatorial complexity that we we can only contribute with pieces of the parcels in a very well defined system that we are able to track structurally and then find some correlates with activity so I think I understand or at least I get the basic idea that the switching is really between just inactive and active when it's repressing but very complicated sets of ligand interactions that combinatorially are controlling that activity correct yeah which seems to be the case in major cases but it has to be very tightly regulated Tim did you want to unmute yourself and ask your question sure and I was wondering about the streptavidin crystals did you ever try processing without subtraction I can imagine that actually because the information is very localised it might not make that much difference have you ever done that comparison you know it just we have not I don't think sometimes we try to do it we don't do it we don't do the removal right and we have issues so that because of that we think that not doing it at all will be even worse but I don't think we've ever tried we have to do this even before we do the movie alignment of the frames okay that's interesting yeah I have two more questions John did you want to ask your question yes that was I thoroughly enjoy that thank you very much this is John DeNu quick question about so in the context of your di-nucleosome do you have a sense of the relative affinities between say the weakest complex binding to the tightest complex binding that overall affinity and I'm asking that because how do you envision spreading occurring in other words what's the binding what's the binding or sorry the thermodynamic aspect of this for release you know what's the release of this guide to actually do this spread I can see the rewrite mechanism but do you have could you paint a little picture about that right so we don't have any yet any direct information on on and off rates on mononucleosome di-nucleosomes nucleosome with different modifications the on and off of the nucleosome versus the tail things that we don't but I can tell you from the failure of our cross-linking experiment that these complexes are not very stably bound to the nucleosomes that they're coming on and off that's why the competition can occur through through the cross-linking even when we have preformed complexes it makes sense no enzyme ever wants to buy it and grab onto onto the substrate so so what I envision these these complexes are constantly coming on and off but what we can tell is that the complex has the capacity to engage nucleosomes that are consecutive to one another if they happen to be in that encounter occurs so that the active that the the amortified complex amortified sorry nucleosome is engaged by the allosteric site that will that will activate and allow him to to act to modify the amortified complex that is right after it also tells you that binding to one nucleosome will favor binding to the other so binding to the allosteric will favor because of the geometry binding to the substrate bind them to the substrate favors binding to the to the one next to it it is modified especially if it's modified so so in those molecular encounters that have to always happen that the amortified complex and the enzyme and its and its substrate there is the possibility of doing two contacts that are geometrically favoring each other in that respect but also that result to activation so that's how you know I think the the spreading is favoring two ways you can you can this has to be better than just adding modified tail to a mononucleosome right because it's not just the act the allosteric activation of the active site is the geometry favors the engagement of the substrate when the other one is engaged to so there's a contribution that comes from the other proximity of the tail but also the sheer geometry of engaging in the context of linker lens that are of the type that we see in our genome okay so so there is a match in the dimensions of the two binding sites for the two nucleosomes and what the linker DNA tend to be in our genome so that's that's how I I envision having this double effect that will favor the modification of a nucleosome that is next to one that has already been modified I don't know if you're very convinced I cannot see your face you know it still comes to release right so you know so what is the binding energy for release so the one thing that we cannot get so one of the things you have to realize is that when we freeze our sample you know there are some complexes that are without nucleosomes and nucleosomes that are by themselves in addition to the complex this is an equilibrium we don't look at the ones that are just nucleosomes because the nucleosome structure is known we don't look at the ps2 because we've already solved that one we concentrate on this but that's an equilibrium we try to purify the ones that are bound with the washes that we found the grid this is an equilibrium this thing is coming on and off I don't have rates the rates we have to do different type of experiments and there are rates that are that go beyond engaging the nucleosome have to do with engagement of the tail and then the catalysis of the reaction and these things how do they match how much once you grab the nucleosome how many opportunities you have to how many during the on on time of the tail can you modify or not and then it's the mono die and trimethylation which are three events that have very different kinetics we have not looked at the kinetics yet we have a structural framework to start thinking about it but we have not done kinetics yet absolutely thank you this has been a great discussion unfortunately we have to rush Ava into another meeting so let's thank Ava again for an outstanding presentation if you did have other questions I'm sure Ava would be happy to answer those by email so thank you again Ava and thank you everybody for coming thank you all for having me