 I belong to a large experimental lab at a higher university. Directed by Karina, I'm a lab theoretician, so I very closely sit with a lot of data produced by different experimental techniques. And I was thinking for years of all the problems that Ellen Reinh just talked about. And I'm going to tell you about two of those problems that come at different stages of HIV viral life cycles. So this whole lab is doing molecular biology of HIV. And I was able to discuss these ideas with Robin Brunsma, and I'm very grateful for this possibility. But I'm going to present the naive first version of it. It's unpublished. It's idea about how HIV is packaged in its own genome. And then there will be probably ramifications on that. And I'm also grateful to be pleased after Professor Ellen Reinh talking, because a lot of what I need to introduce to be able to explain these ideas were already introduced by Ellen. So among the all possible viruses and how they live and perform their basic function of promoting their own genomes that Shura Grosberg was talking about, the retroviruses are a very special niche. So of course, just like every other viruses, they need to pack their genome. They need to go out of the cell. And then they need to go and infect another cell. And at this stage, they need to encode and to make their genomes available for replication and for translation to make their own viral proteins. And as we know, it's basically, as Shura was saying, three components. The nucleic acid, that is the genome, the capsaid that forms the cell, and probably the membrane in a lot of cases. And retroviruses, they all solve this problem a little differently. But the basic thing is that the stability at the stage of assembly comes from the membrane. Those are enveloped viruses. And the assembly, as was discussed here many times. So there are two different scenarios that were discussed here generally, right? One scenario of assembly is when the capsid proteins are interacting with each other very strongly, and they assemble on their own, maybe with a little help of nucleic acid that they package. Sometimes even they don't need nucleic acid. And the nucleic acid is packaged after or even needs energy to be packed in, right? This is one class. The other class is all that all retroviruses belong to is when the interaction with the genome, with the nucleic acid, is much, much stronger than the capsid-capsid interactions. And the assembly never happens without the genome, no matter how high salt or what pH you're using. And the capsid-capsid interactions are a small addition on top of this major very strong interaction with nucleic acid. And the HIV is a very good example of that, and we're going to talk about that. And in fact, I believe that assembly, some sort of assembly happens even when there are no gag-gag interactions whatsoever. So after, so the other step of like encoding happens to a completely different capsid, right? There are two different capsids. One is immature capsid before the gag is processed and cut into pieces. And the other is mature capsid. That is this cone that we've seen several times throughout this meeting. And my second story, if I have time, is going to be about the step of encoding. This is a very interesting story that Robin and I published on about three years ago. But since then, there was a lot of experimental interest. And I think that basically confirms this idea. And I brought a poster about that, but I forgot to put it up. So I put it as part of my talk. So OK. So the problems of selective packaging of HIV genome was pretty much presented by Ellen very well. So the problem is that we have that this gag protein that self-assembling without the help of anything else, in principle. Maybe the membrane or the negatively charged little molecules, AP6, that Ellen was talking about, that can substitute for the membrane. So it assembles. And it assembles just as well on any random RNA non-selectively. No difference in the virion assembly, even in kinetics, maybe in the lag time before assembly, if at all. And in vitro, in solution, and even in vivo, there is no difference in the binding strength of the size sequence that is a packaging sequence in genomic RNA versus any random piece of RNA. And there is a huge access of this mRNAs that Ellen introduced are packaged non-selectively in these genomes. And at the same time, the packaging selectivity for the genomic RNA is pretty high, 80%. 60% or 80%. It's not 100%, but it's very significant. So the other problem of there seem to be conditions when there is enough gag in the cytoplasm, and you can see it by fluorescence. And it's about in equal amounts in the cytoplasm and on the plasma membrane sitting there doing nothing. There is no assembly. There seem to be a critical concentration of gag that is needed to initiate any sort of assembly. And the very late kind of very recent studies from Wysze, whose lab at NIH, show that if we have a genomic RNA, then at the same low gag concentration, the assembly can be done. And there is a virion production that is much more efficient and significant. But if there is no genomic RNA, then at the same gag concentration, there is no assembly at all. And then there is this mysterious rule of genome dimerization. We know that genome is packaged as a dimer. In this little sequence that is called dimerization domain of genomic RNA, there are complementary base pairs in the loop, just three, four base pairs that are absolutely critical for dimer formation of this genome. And this is at the end of the RNA. And this dimerization signal is within the Cy region, which is the packaging region. So it's naturally to kind of speculate that the dimerization being in the same exact place as the specific region of genome important for packaging, that that's what's important, that they're close together and that they're forming the same specific site for assembly. But again, experimentally, when people measure in the test tube, if binding to the dimer RNA that is dimerized is stronger than to the monomer that was mutated in this stem loop and precludes dimerization, they don't see much difference for dimer affinity versus monomer affinity. So this is the crucial experiment that was done fairly recently in vivo in cells that I think that makes us believe that the selectivity for the genomic RNA comes at the step of assembly nucleation. So this part, so what is being done here, it is in vivo. People are taking, this is done by Sabla Kutli and Paul and they are taking this cytoplasmic fraction of cell and they're measuring, they're quantifying by RNA PCR. What kind of RNAs are in this fraction available for sequencing essentially? And they see that there is very little of the viral RNA in the cytoplasm, just as you would expect it should be. Because viral RNA is one of the, just like any other mRNA, but it's just one sort of mRNA. It's a product of one gene that was permanently introduced into the genome. So the amount of viral RNA is maybe 10th of a percent or maybe hundreds of a percent. It was detected, but you can't even tell here how little it is. And 80% is some cellular random RNAs, right? But then they do the same cross-linking it to GAG and watching what fraction of GAG cross-link to viral RNA as opposed to cellular RNA. And they see that about, like, say, on average, 3% of the GAG cross-link to the viral RNA. 3% is more than, like, say, 0.1%, right? So there is some level of selectivity at the stage of cytoplasmic binding of GAG to genomic RNA versus random RNA, right? But this is certainly not enough to selectively package, right? This is the degree of selectivity in the cytoplasm. It is maybe, I would say, within 10-fold, less than 10-fold, maybe 3-fold, right? However, when they do the same two experiments inside the virions, so they purify the virions and they analyze in the same way the content of this virions. What is the RNA in this virions? Non-specifically. Now, the cellular RNA constitutes about, like, say, 20%. It's non-zero. It's significant. There are some cellular RNAs being packaged in the virion. And about 60% to 80% is viral RNA. So the enrichment happens somewhere in between cytoplasmic binding of GAG and complete virion formation, somewhere in between. So it gives us an idea that, probably, most likely, it's a nucleation step. Because the rest of the RNA, besides cy, is just as good as any random RNA. There is no anything specific about that RNA that could confer the selectivity of packaging. So in the same experiment by Ceble and Paul Binache, of all of the GAGs that cross-linked to the viral RNA that was this very small percent in the cytoplasm, like, say, 3%, right? They were able to see what were the specific binding sites on the dynamic RNA that the GAG cross-linked to. And they specifically have shown that these are the three sites in the cy region, right? It's very kind of gratifying to see after everything we've heard from Ellen just now, that there are these three strong sites in the cy signal for GAG or for a nucleic acid. And they were able to map out those sites pretty well. And this mapping coincides very well with the mapping that we've heard about by shape analysis from Ellen just now. And also that comes from Carine Museum's foresight lab. We do it by this salt-efficient binding studies. And we see the same pretty much specific sites in the cy. And then in our lab, they've done three-dimensional sacks of this cy, right? It's a pretty rigid, flat structure. And there is a three-way junction in the middle. And those three sites map very well to this three-way junction, right? By the place where three storms come out. So we'll come back to the importance of these three sites and what can it mean? Finally, we have to consider all different possibilities for selective genomic RNA packaging because it's all in vivo and in the cells. And people can't see the specificity of genomic RNA binding to GAG. They, of course, infer a lot of other proteins that are cellular proteins that can be important for the selective packaging. Some motors that drive and collect GAG in some place and not the other or anything you can imagine. There are million possibilities how the selective packaging can happen. We know it doesn't happen co-translationally that the mRNA that produces GAG doesn't get packaged. We also know that there is no compartmentalization of the genomic RNA with the GAG in the cell. But this is a very simple, as simple as it can get in vitro experiment that was able to reproduce in some way the selectivity of genomic RNA packaging. This is in vitro selective packaging assay that was just recently done in Jim Crowley's lab. And what is being done here, it is a giant unilamellar vesicle that in its content reproduces the plasma membrane composition. So it has these peeps that are highly negatively charged, small molecules. And it has a wide composition of the negatively charged polar heads and some, a little bit of cholesterol. So it is pretty much like a plasma membrane, right? And it is in red here. And then fluorescently labeled in white is GAG. And this GAG is seen as a puncture here. And then co-localizing with this white panktole of GAG, there is a green genomic RNA fluorescently labeled, right? And what they show is that they fluorescently label different RNAs that are of different degree of specificity. This site-containing region from genomic RNA is labeled or completely random, no, not quite random. It's another region of the genome that was cross-linking to GAG, just a little worse than the genomic, than the site region. And then there is a completely random piece of the same length. It's not a full length. RNA is just about 370 nucleotide. And what they were able to show that there is a complete co-localization of GAG puncture with the genomic RNA, even when there is a tiny little bit of genomic RNA, like 0.5 nanomolar. But if the, and that happens with any RNA, it's completely co-localizes with a puncture. But when they start adding 10-fold access of genomic RNA, the, the, so the genomic RNA is able to withstand competition with the non-genomic RNA, no matter how hard, no, even 10-fold access of non-genomic RNA. But if it is a specific RNA that is competing, which is in this case, then it is the fluorescently labeled genomic RNA is all competed out by it. So this is what I'm presenting this data because I want to show that the physics attempt to try to understand the selectivity of packaging involving just a membrane RNA. And the GAG is not hopeless that it is being done in the lab that apparently you don't need any cellular proteins to do that. So this is a fluorescent image of the, so this is a low expression of GAG. And the GAG is about half and half kind of dispersed in the cytoplasm and dispersed on the membrane. When the GAG hits a level of expression, then most of the GAG is in form of the pretty much cluster dense form on the membrane and there is none of it in the cytoplasm. It's a pretty critical phenomenon. So then about the assembly kinetics at all. So what people know about assembly is that the genomic RNA is being picked up in the cytoplasm by very few GAGs that people can't see fluorescently because it's just, it's below 10 GAGs together that do the selection in the cytoplasm. And this is the slow selectivity step in the cytoplasm that we talked about. And then the first thing, so they watched this assembly by turf and turf only observes the fluorescence from the membrane, right? It's a two-dimensional thing close to the membrane. So what people see, the first see, the green is the genomic RNA coming to the cytoplasm. It comes pretty abruptly and it probably is brought by a few GAGs that people can see, right? And then over a pretty long time, which is 10 to 20 minutes, there is association, gradual association of the GAG with this nucleus that is being formed on the membrane. So this is a very slow time that is definitely not diffusion limited and controlled by something completely different. And we also don't see a very long log time, like a lag time that happens before this first genomic RNA that we can see on the membrane. It's a really long time and sometimes it's infinite in terms of if it's a low GAG concentration, the GAG sits forever before nucleating the assembly, right? So it was also shown by Weisha, whose lab that all of the GAG that's gradually accumulates here in the virial comes not from the membrane. There is tons of GAG sitting on the membrane and you would think that it would be easier to diffuse to this nucleus right from the membrane because diffusion in the membrane is maybe 10 times slower than in the cytoplasm. But it doesn't come from the cytoplasm. All of the GAG that is assembling in this virions over this 10 minutes comes from the cytoplasm. So to be able to move any further in this kind of thinking about what selects the genomic RNA, we have to try to see what the GAG is, right? So as Ellen already introduced, it's made of cellular globular domains, right? And it's n-terminal and it's c-terminal are the globular domains of matrix that is supposed to bind to the membrane in immature assembled virion. And nucleic acid that is supposed to bind to the RNA inside the virion. And from mostly studies from Ellen's lab, we know that this individual GAG is highly flexible. In particular, this linker between the n-terminal of capsid and the matrix is 30 amino acid long, completely flexible unstructured region that is just as good as completely random polymer with a persistent length of one amino acid. So this matrix domain has no problem reaching anywhere in the volume of this protein. So what we also know is that both matrix domain and nucleic acid domain by either RNA or plasma membrane electrostatically, competitively, and with about the same affinity. Nuclear capsid binds plasma membrane just as well as RNA, maybe a little better for specific RNA. And matrix domain binds any RNA non-specifically very strongly and almost as strongly as the plasma membrane, but if it has this meristel tail, which is hydrophobic tail on it, it's meristelated, cotranslationally at its n-terminal. Then this meristel allows to bind matrix to the plasma membrane about 10-fold stronger. This 10-fold stronger is very essential, but the 10-fold stronger in terms of the binding free energy gives only the logarithm of 10, which is about 2KT. So it's stronger, but not that much stronger. So there is always this competition between Gag binding to RNA in the cytoplasm and Gag binding to the membrane, right? And the difference in terms of the real binding energy is just this minor preference, essentially, of matrix to membrane because of this hydrocarbon tail. So now there is a statement here that is kind of controversial so that the Gag-Gag interactions in immature assembly are weak and this, so we just heard from Ellen how important they could be, right? And there is no real number placed on how important, how weak or how strong they are. What I was able to estimate on all of the binding data, Ellen's and from our lab and from any lab I could get to, right, is about 2KT. So the capsid-capsid interactions are not very strong, including the SP1 interaction when it's turned on because of many Gags are together. It's just like another contribution to the short-range Gag-Gag interaction. There is also interesting observation that comes from Ellen's lab that if he mutates this major dimerization domain of Gag in the capsid WM site or if he mutates the SP1 region, which is another contribution to Gag-Gag interaction, there is still assembly. So this is a well-typed assembly. These are this nice little 100 nanometer invariance. And this is the assembly with the major dimerization site of Gag-muted. There is some macroscopic structures that contain membrane dense layer of Gag and RNA in them. They may be not perfect invariance. They don't retain the same curvature, but there are still some sort of assemblies with the same components, almost as dense as the normal invariance. So the assembly happens, you can see, without Gag-Gag interaction, which is very strange. So I'm trying to make as simple model as I can of this what happens. And I'm starting with trying to consider possible binding states of Gag. So the free Gag, I'm trying to think of the Gag as matrix and C. There are both cationic domains connected by a flexible linker. Okay, it's kind of simplistic. And by doing that, I'm neglecting, for now I will then introduce it, the Gag-Gag interactions. So the interactions in physiological salt, 150 millimolar of this particular model for Gag with either RNA or membrane are very, very strong. It's 20 kT's. And about 10 kT's comes from matrix and 10 kT's comes from a nuclear capsid. So, and they are both binding of course at the same time, right? It's the same molecule and very strong two interactions. And these interactions I'm taking from measured kT's of individual NC domains and matrix domains measured how they bind to the membrane or how they bind to the RNA, right? And then there is a third bound state of Gag that is in between matrix, in between the plasma membrane and RNA. And it has also the same binding strength here and here. So there is no free Gag floating anywhere, right? And everything that's on the membrane is either in that state or in that state. So this state can't assemble and that state can't assemble. But this state is a nucleus for assembly, even one. All we need for assembly is three things together, the membrane, the linker, which is even one Gag and the RNA. So we can now, because we know the binding free energies of matrix and NC to each either membrane or RNA, the free energy of binding of like say this Gag to a membrane is this free energies of interaction of individual domains and the entropy cost of localizing the Gag, right? Which depends on Gag concentration in the site of plasma. And then the free energy of binding this Gag to the RNA is the same terms of interaction with the RNA. And now it's an entropic cost of localizing this RNA that is in excess over Gag, right? And when we form a triple complex, this complex is very strongly disfavored by the fact that we need to localize both Gag and RNA. So it's disfavored by both of these terms, which are huge. And these terms are on the same order of that, right? But now let's consider the transitions between the states, right? The free energy differences between Gag bound state to RNA and the extended state. So I want to see how probable is that state because we are interested in the state as a nucleus for assembly nucleation, right? So these transitions are really driven by a stronger binding of like say in this case, nuclear capsid to RNA as compared to the, so I'm sorry, this is transition from RNA bound state to RNA and membrane bound. So this is driven by matrix binding a bit stronger to the membrane than to the RNA, but it's disfavored by the necessity to localize RNA by the membrane, right? So this difference is 2KT and this entropy loss is 10KT. So this is very unfavorable process. And the same is if we want to unbind the RNA, the nuclear capsid from the membrane and rebind it to RNA. This is about at most like 1KT, 1-2KT process that drives it and the entropy loss is again 10KT is huge. I'm constructing a phase diagram for a single Gag, right? So these are these free energy differences for these two transitions from the RNA bound Gag to extended Gag like this one. And this is the free energy difference between this band Gag and plasma membrane to extended Gag, right? And physiologically, we're always sitting in this kind of situation. When we don't have extended Gag at all, it's either all bent on RNA or all bent on the membrane, right? This situation is highly unfavorable. And then there can be some cases here. So it all depends on the concentration of Gag and RNA, but physiologically, we're always in this regime, way below. And this is what we need to start the nucleation. So it doesn't happen for one Gag. One Gag can do that. So now if we imagine that we have a special RNA that binds not just one, but like say two or maybe three Gags in the same place, not because they're interacting with each other, but simply because it has a slightly stronger binding sites on this RNA that are all close together sitting here. Now this process of this Gags extending and binding to the membrane has the benefit of three mate vases slightly stronger binding to the membrane, which makes it instead of two KTs, six KTs for three Gags, right? And we're getting it at a price of localizing just one RNA, not three RNAs, right? Which makes it much more favorable. So this extended state is much more favorable than just one Gag. One Gag can do that, but two and better three probably can do that. So this is good, but, and these are like coming back to this importance of the three sites together on the site sequence. Now why is it, so if we are able to nucleate this nucleus, then the assembly is driven and it's very strongly driven even in the absence of capsid-capsid interactions, Gag-Gag interactions, why is it driven? Because after the structure is made for whatever reason, if there is a link between the membrane negatively charged membrane and negatively charged long RNA, it's as careful for the future assembly. The rest of the Gags that are sitting in the cytoplasm bound to some random solar RNAs as monuments, they're joining this assembly simply because they want to release the RNA that they're bound to. And this release of the RNA costs 10 KT of entropy that is liberated as the single Gag releases the RNA and joins the scuffle that is already there, pre-made, right? And that drives the assembly pretty strongly without any other components. So this is how the assembly is driven and it's essentially the ground state is a macroscopic assembly that is for me. So these are my conclusions that are, how many minutes do I have? Then self, yeah. So I pretty much already gone through all of that so I don't wanna waste time. And I would like to tell you a completely different story about completely different stage of the virus life cycle. This is about unquoting of this completely different capsules, right? So far we were talking about formation of this immature capsules, right? That is basically a layer of the membrane and a layer of the RNA and the Gag in between, right? And now we're talking about this pretty well-organized crystalline structure that you've seen several times through this meeting already. It's a conical shaped capsid. And we know it's of very low stability because people who are trying to purify it and study it in labs were having trouble with it for many, many years because it's a very low stability structure. But you can make a structure of it without purifying it inside the membrane and you can do cryaminate and all sorts of structural studies and the structure of it by now is known with the like say three angstrom resolution. It's a most well-defined structure. It was very well modeled by Klaus Schulman, Schulten's lab. And the question that was there for years, so like about maybe five years ago people would always say and it was written in textbooks that this structure as it gets into the cytoplasm of infected cell that it immediately falls apart. And the reverse transcription that falls happens in the cytoplasm of infected cell. And there was no question about that. It was kind of a textbook notion. So about five, six years ago people started to question that. And to me it was also over the puzzle when I learned about it because what makes the reverse transcription is a very special polymerase that is a viral protein that is completely absent in any sort of mammalian cell. It's a reverse transcriptase that uses the RNA as a template to make DNA. And also it can use the same polymerase uses the single-stranded DNA to make another DNA to make a double-stranded DNA, proviral DNA. So there are just 200 copies of this tRNA inside this mature capsid. And if mature capsid is gone, then this RT will be also gone because it's an unspecifically binding, nucleic acid binding protein that can bind to any RNA in the cytoplasm. So there will be no reverse transcription if the encoding happens prior to reverse transcription. So I would expect that the capsid should stay almost till the end of reverse transcription. But it was not known. So this is the typical picture that the infection happens. The capsid falls apart. The reverse transcription happens. The double-stranded DNA is then made in the cytoplasm and then somehow transported through the nuclear pore and gets inserted into the human genome. The later picture that comes from the studies starting like 2009 is that the mature core can survive all the way through reverse transcription, can come to the nuclear pore. And at this point, there is an encoding happening. And the DNA is somehow pushed through the pore and that the capsid even can help to localize the double-stranded DNA to the pore and to transport it inside. So just to remind you that these are two completely different immature and mature variants. And what happens here is that the gag is processed in the components. And what assembles in the mature capsid is just the capsid. The matrix stays by the membrane. The nuclear capsid that is bound to RNA is processed from capsid and aggregates with the nucleic acid with RNA inside this capsid. And it takes up a very small volume of this capsid. The RNA condensed by a nuclear capsid is a small little dark spot on this TM images in this mature capsid. So there is no problem of packaging this viral RNA inside mature capsid because of the nuclear capsid because it's a very strongly condensing agent. And the way it condenses the nucleic acid, just coming back to that, the way that nuclear capsid condenses nucleic acid is pretty much like multivalent cations can condense any sort of nucleic acid. I believe it's a counter-incorrelation mechanism that a lot of people in this room had worked on before. Just like cobalt hexamin or spermidin or spermin would condense any sort of nucleic acid. It's a multivalent cation with a charge 3 and 1 half nuclear capsid protein. So now there was a notion for a very long time that this capsid should be good for reverse transcription. There are about 10 to 8 nanometer holes in this structure that are transparent to nucleotides and say to nuclear transcription inhibitors, but they're not transparent to any other larger molecules. For example, even to the smallest protein like nuclear capsid, 55 MNS. But the reverse transcription was observed for years in dodging this reverse transcription. People had this cone inside the membrane, as it is purified in purified virions. And people permeabilized this membrane with a detergent and add nucleotides. And the reverse transcription goes just fine, no problem. There are no enzymes, no cellular proteins, no nothing needed for reverse transcription. And lately, just recently, one month ago, I have heard of people observing the complete reverse transcription now in the virions in vivo in mature capsids without shells. So it definitely happens. So also, there is just a reason this year publication in science that shows that this little pores, 8 nanometer pores in the mature capsid, are not only transparent to nucleotides, but they're actually actively transporting nucleotides. They're made to transport the nucleotides inside. And it is a very efficient, fast process. And the slight mutations, it's just made, I don't think it consumes the energy, but they're made of five arginines, positively charged sitting right here, that are very efficiently binding, and then releasing inside. It's made, kind of, it's designed, right? And very small mutations in this pore make a huge difference on the rate of reverse transcription. So there were early studies of the interrelationship between the reverse transcription that might happen inside this capsid and the rate of encoding as it was observed in vivo. So the rate of encoding was quantified by measuring the amount of capsid in the capsid as a function of time post infection. And the encoding happens about an hour after infection. So the capsid's fall apart. They don't know if they fall apart at once or gradually, they just measure how they fall apart over time post infection. And if they halt the reverse transcription with Neuropean, which is the reverse transcription inhibitor, then the encoding just doesn't happen. The cone, this mature cone sits there forever. And at the same time, they're monitoring the progress of reverse transcription. And in this early study, it wasn't the reverse transcription and the same capsids that were falling apart, just the rate of reverse transcription post infection. And the reverse transcription products that happen, the late reverse transcription, which means double stranded DNA, appeared about the same time that the encoding happens. And that is a link between the reverse transcription and the encoding. They have the same time. And also, if the capsid is stabilized, so the unstable capsid falls apart much faster. And the stable capsid survives much longer. And if the capsid stability is modified, like by capsid mutations, artificial capsid mutations, or by changing the host cell proteins that stabilize or destabilize this capsid, binding to it in the cytoplasm, then the infectivity of the virus is affected like a thousand times, many, many times. It's a huge effect on infectivity. So now, getting back to thinking about reverse transcription happening in this mature capsid. So it is absolutely critical that the nuclear capsid inside this capsid is condensed. The condensing agent is this nuclear capsid protein that Alan talked extensively about, so I don't need to talk about it. What I want to say is that in the salt, salt competition assays, essentially measuring the KD as a function of salt that Alan talked about, it looks just like cation with a charge three and a half. And it's fairly nonspecific. It has some specificity that Alan also talked about that's coming from this aromatic residues sticking with the Gs, but it's fairly moderate specificity. And we know what happens to RNA or DNA if we add to it, like, say, cobalt hexaminase permident. This is the double-stranded DNA condensed with cobalt hexaminase permident. And these are classical experiments and what we know about it is that all of the double-stranded DNA within a second compacts in the two orbital structures, right? I assume that the nuclear capsid does exactly the same to either single-stranded DNA or double-stranded DNA that is being made in course of reverse transcription. So thinking that, Robin and I had made a theory how RNA is being transformed into a double-stranded DNA using this reverse transcription, RIT polymerization. And this is a very rigid structure. This is a very compact structure. The self-volume of completely done DNA is the same as a single-stranded DNA, but it's rigid. So it needs to be compacted by a nuclear capsid protein that is there. If there is a little hole in the capsid, the nuclear capsid will escape and this toroid will decondense and that will be the end of reverse transcription, complete blow-up of the capsid, right? And there is a balance between the self-attraction induced by a nuclear capsid that makes the size of the toroid that is pushing on the capsid. And the elasticity of this capsid. And by considering this two things, elasticity of the capsid and of the toroid, we made a phase diagram for what volume of the double-stranded DNA inside the capsid or what length of the DNA being made inside the capsid can be withstanded by the... So on this axis is a self-attraction of DNA by a nuclear capsid. On this axis is the volume fraction of the double-stranded DNA inside the capsid. And the different lines correspond to different capsid stabilities. So if the capsid has no stability, just upon toroid touching it, it falls apart. That is how much of the DNA it can tolerate being made inside. But if the capsid is more stable, then it can tolerate a little more DNA being made inside. If the capsid is too stable, it will never be unquoted. So just now this year, they appeared the first FM studies of the capsid. So this is mature capsid on the FM grid being scanned by FM. At some point, the reverse transcription is induced by adding nucleotides to it and it's still sitting on this FM grid. And at some point, there is a filament that pushes on the capsid from the inside and that makes this kind of bump in the shape of the mature capsid. And this kind of filament that they can trace with FM appears at the same time that the rigidity of the capsid that they can also measure by FM goes up and then the capsid is kind of, the rigidity goes to zero. And they can also measure that if the capsid is overstabilized by mutating the capsid, then there is a disincrease in the capsid rigidity that corresponds to the filament formation in the capsid. But the capsid never disassembles, it just relaxes to some more or less previous rigidity. Okay, so it looks like this FM data supports our picture of mature capsid being blown up by reverse transcription happening inside it. That's pretty much, thank you.