Scientist Stories: Anne Moscona, Entry of Human Paramyxoviruses

I'm really, really pleased to welcome Ann Moscona today. So viruses are bad. Viruses are awful. And key to any virus, whether it's one that caused a pandemic or ones that might cause the next pandemic or the ones that really irritate you on a day-to-day basis, is that event whereby the viral entry glycoprotein attaches to the human cell and drives itself in. Now even though that's the single fundamental step that decides whether or not the virus is going to infect your cell and whether or not your therapeutic intervention and any B-cell mediated vaccine is going to work, there are still major black boxes there. We know the before fusion. Sometimes we know the after fusion. But what happens in the act of fusion is just a black box. We make these cartoons. We try to model it. We have no idea. Anne's work aims to capture those intermediates and understand how to interfere with virus entry with lots of different kinds of inhibitors and different strategies. And she's gone about it from just a beautiful series of complementary techniques. So she comes from a scientist family, or both her parents were developmental biologists. She got her bachelor's degree at Harvard and her medical degree at Columbia. Her laboratory is now at Columbia. And it brings together molecular biology, virology, immunology, biophysics, computational biology, and structural biology to understand viral entry. And although she's done a lot of work on SARS-CoV-2, most of her body of work has been on a family of viruses called the paramyxoviruses. And my money around the paramyxoviruses is the sleeper family of viruses that might cause the next pandemic. Because in that family are ones with an R-naught of 18, and ones that are 90% lethal. Also in there are ones that you have been up all night with your kids when they've had croup. That's parainfluenza virus or respiratory syncytial virus. So these are medically important viruses, and her work has just, has beautifully illuminated how they enter and how to defend. She is a fellow of the American Academy of Microbiology, and she is the next president of the American Society for Virology. And I'm so pleased to have you here today.

Thank you, Erica. I'm so excited to be here and get to see this place. And Erica just described my work better than I possibly can, so thank you so much for that introduction. Since it's a small group, I hope that you'll be really interactive and butt in and ask me questions. So, um, we're off mute and we're showing the video to Zoom, right?

Okay.

So as Erica said, paramyxoviruses are really important, and I came at this from thinking about children initially, and paramyxoviruses are a major cause of child morbidity and mortality worldwide. In fact, respiratory viruses at the beginning of when I started working in this field overtook diarrhea as a cause of death of children under 5. And really the bulk of these respiratory illnesses are caused by viruses. Parainfluenza, respiratory syncytial virus, and metapneumovirus are all paramyxoviruses, although RSV has been regrouped into respiroviruses. And of course influenza. And they really take a major toll in terms of child health. And of course also measles is included in that, but even outside that group. So we add measles to the serious paramyxoviruses and what Erica alluded to, Nipah and Hendra emerging. Do I need to do something else? Zoom does not hear, sir.

Apologize for—

yeah. Nipah and Hendra are the ones that Erica referred to as being terribly lethal and potentially emerging, and measles has the R-naught of 18 and is incredibly transmissible. So this is a really, a really important combination of traits that these paramyxoviruses have. It looks like— Those look on, right? Yeah. Okay, good. So paramyxoviruses have a lot of similarities, and I'm going to tell you the general themes about how they enter cells, and then we're going to get more specific.

-specific.

So paramyxoviruses are all negative-sense RNA viruses. Their genome is a single long strand of negative-sense RNA and they are all enveloped. And— oops, sorry. And enveloped with a lipid bilayer that they derive from the host cell. And into this lipid bilayer envelope are inserted the surface glycoproteins of the virus. A receptor binding protein and a fusion protein. And the first step in entry is that the receptor binding protein will bind to a target cell receptor, which is different among different paramyxoviruses. The receptor-binding protein then actively triggers the fusion process. Virus enters the cell, undergoes its replication, in general sense, entirely in the cytoplasm, and then buds through an area of the membrane where the glycoproteins have been inserted, aided in many interesting ways that you know a lot about, by the matrix protein and ultimately bud to produce new virus. So just to zoom in on that process, again, receptor binding occurs, actively triggers the fusion protein to insert into the target cell, and then to initiate membrane merger and fusion. And for the virus that I'm really going to focus on, which has been sort of the prototype example of a lot of the work that I've done, We're going to talk about parainfluenza virus for which the receptor binding protein is hemagglutinin neuraminidase or HN and the fusion protein is F. And that's the case for all of them. So the hemagglutinin neuraminidase or HN binds to the receptor, to the receptor, actively triggers the fusion protein and mediates fusion. And when I entered the field, it was thought that the two proteins function completely separately. Receptor binding protein binds, fusion protein fuses. All you need for fusion is the fusion protein. And that information came from studying things in isolation, studying expressed proteins in isolation, studying structures in isolation. And really, it was not until we put things together and tried to see what actually happens in real life that we realized that the Hn receptor-binding protein actively is in a complex with the fusion protein, triggers the fusion protein, and then the whole thing works as a little molecular machine that I'm going to focus on. So just zooming in further on the steps of entry, up here is representing the host cell with, in the case of parainfluenza, sialic acid-containing receptor. And here are the Hn receptor-binding protein and F prior to engagement. And the work that we've done in the last, really, that's come out in the last few months to year, a lot of my old schematics like this are sort of wrong. I'm going to have to remake all my introductory slides. And I'm going to show you that in the end. Okay, so prior to engagement with sialic acid receptor, the HN and the F are together. And F has to be cleaved. Activated prior to being— prior to the virus being infectious. And there's— it's made as a proprotein and then cleaved into F1 and F2 that stay disulfide-linked. And so then HN engages sialic acid receptor and then triggers the fusion protein to undergo this large conformational change and insert its hydrophobic tip into the target membrane. And then it folds back on itself, the HLA heptad repeat domains at the N-terminus and C-terminus want to meet each other and they're driven to come together and that sort of drives the fusion of the viral and the host membranes. So we really have focused on is this, how does the receptor engagement by the HN, receptor engagement by the receptor binding protein cause fusion to get triggered? What is that process and how does it work? And, but as you can imagine, I'm going to keep coming back to this, all these steps are potential targets for antiviral intervention. So HN and F, this is just a sort of a cartoon based on previous models made from crystal structures of H, of Paraflu, and SB5, Paraflu5F. And H with the stalks kind of drawn in this dotted lines, they're together on the surface. Of the virion. And looking down, just, this is sort of a whirlwind zoom through HN and F. Looking down on the globular heads of HN, and please just tell me if I lose you at any point because I just want everyone to stay with me. But looking down on the globular heads, there's the primary binding site, or what we call site 1, which is where HN globular heads bind to sialic acid receptor. And I'm going to refer to that all the time as the receptor binding site 1. What we found is that at the dimer interface between the globular heads, there's a site that we call site 2, which is also important in receptor binding site and important for activation of F. We'll get back to that. F has been crystallized, had been crystallized in both its prefusion, so the before receptor engagement conformation, and then its post-fusion post-fusion state, which is after it's already extended and then refolded on itself to mediate membrane merger. But of course, nothing in between, no intermediates had ever been seen, just the sort of like what Erica mentioned, just before and then after. So what we had shown over a bunch of years is that interaction between HN and F, between the receptor binding protein and the fusion protein, before and during receptor engagement completely regulates modulates the ability of viruses to fuse and enter cells. If those proteins aren't interacting or if their interaction is different, it modulates the ability of the virus to enter. And a lot of strategies that we use to figure that out and to understand how HN and F are interacting were based on mutations in HN that arose under selective pressure. So remember, coming into this, not even knowing that HN affected fusion, we learned that mutations in the primary binding site or in the secondary binding site, site 1 or site 2, that increased receptor binding increased fusion of the virus into the cell. Mutations at that site 2, that dimer interface site, that affected F triggering, so literally one amino acid residue in the dimer interface greatly enhanced the ability of HN to activate F and activate fusion. Fusion, those would greatly increase the fusion of the virus. In contrast to mutation in the stock of HN that decreased F triggering would decrease fusion. So functions of HN with F being completely wild type, individual residues in HN would massively affect viral entry and the ability of HN to trigger F. So those site 2 mutations, for example, mutations in the dimer interface, individual mutations in HN, when you look at— when you use a cell fusion reporter gene assay to look at the ability of HN and F to mediate fusion, these single amino acid changes at site 2 massively increase F triggering, massively increase fusion over time, and massively increase cell-cell fusion in this assay. So we basically had these super fusogenic viruses based on individual mutations at the dimer interface of HN. Everything else in the virus being the same. And we were like, they would eat through a monolayer. They were just these really great viruses. So then we put them into lung airway epithelium, these red viruses, these super effusers, and they didn't grow at all. They didn't grow in the airway. They didn't grow in the lung. They were super, super effusers in cells. Great in vitro. Great in culture. Don't work in the lung. Don't work in the airway. But under selective pressure, which is a way that we really let the virus, biology of the virus tell us what's happening in the mechanism under selective pressure to grow in the airway, mutations emerged. This superfuser in the dish had to grow in the airway. Mutations emerged, and the first mutation that occurred was an F right here, and it made this F shy, really hard to trigger. So it was made— immediately the first thing it tried to do desperately was to become less triggerable, less activatable. And the second mutation that sort of fixed in this virus was a mutation in that dimer interface right near the original superfuser mutation. We got a second mutation that made it less actively triggering, less powerful at activating F. So in order to grow in the real world in the long run in animal, you had to pull back triggering, pull back HN's ability to activate, and go back to that balance of just right activation to grow. And we asked whether these compensatory mutations affected HNF interaction in the complex, and of course they did. They modulated the interaction, the intensity of interaction between HNF in that complex. So then finally we asked, after this selective pressure of a superfusing virus to grow in an airway with mutations in the dimer interface? Does this look just like a virus that grows in the field? And yes, it did. The properties of fusion, fusion triggering, interaction between the HNF and the complex receptor avidity look just like the clinical isolate. So we kind of evolved properties that led us to understand what fitness might look like in an actual clinical isolate. And this time The idea that the actual HNF complex and the properties of binding, triggering, fusion really were finely tuned to particular tissues, to growing in an airway, was really, really exciting and kind of influenced how we looked at everything since then. We look at clinical viruses that actually are fit for growth. We look at airway tissues from humans. And really try to understand things in an authentic setting, as I'll show you more. So—

So were those, the mutants and the clinical isolates, similar in terms of the in vitro fusion assay, or were they much weaker in terms of the—

Much weaker. So viruses that are actually fit to grow in a human are really lame in a dish.— and converts.

That's a problem when you try to understand it.

Yeah, it makes it harder, but worth it. I'll show you.

So if the fusion was further decreased, would it make it even more infectious? Or you need fusion, right?

Yeah, you need fusion.

If you remove fusion altogether, it doesn't work.

It has to be just the right balance. Right, yeah, you absolutely need fusion. Look, I'll show you more. So basically, yeah, so triggering a fusion must be just right. So this led us to our— this led us to a concept about a potential target for antiviral. If we tickle the site 2 and make fusion triggered before, you know, prematurely, before it should, make it too much. We kind of copy, copy the phenotype of that superfuser and it could be an antiviral target. So the idea is if you, in the normal process, is what I've shown you up here, but if you tickle that HN to activate F prematurely, you could inactivate virus. And in fact, we've done that mimicking that overactive triggering handicap of the superfusers. And in collaboration with Nir Bental at Tel Aviv University, at Duke University, we designed this small molecule which interacts with the second binding site/triggering site at site 2 in the HN dimer interface. And in fact, it induces conformational change in F by interacting with HN. So it induces HN to prematurely trigger F, and we can see this by when F becomes triggered, it becomes protease sensitive. And after we use this compound, we instigate protease sensitivity of F. And also it works in the airway epithelium and activates virus and reduces replication in the airway. And when we use this in an animal— and this is just a proof of concept— we can use this in an animal to inactivate virus and get, and for some of these sort of prototype compounds, we can get reduction of about a log in titer of virus in lung tissue. And I have actually a cryo-EM showing the activation of F in the presence of the compound, but only when HN is present. But this is like a year or two ago, and it's so low resolution compared to what we have now that I'm embarrassed to show it, so I'm not going to show it, but you can also— I had that slide and I was like, no, I'm not going to show that. But you can actually really see by cryo-EM that F is activated by a compound that interacts only with HN. So yeah, so just binding of this receptor analog to HN activates F and you inactivate virus. And this is a proof of concept. It's nothing that's good enough to like develop yet, but we think it's an interesting way to inactivate virus. Okay. So maybe I'm missing it.

Does hN actually bind physically to F or not?

Well, all we knew at this point by our biochemical, biological, cell-based assays was that it is present, that it has to be present, and that it interacts differently depending on mutations at different sites. So we hypothesized for a long time, but I'm going to show you today, that there is— that it is present in a complex together with hnRNAs. But nobody had ever, as I'm going to talk about, had ever been able to show that. And we thought that based on whole decades of biochemical data. [Speaker:Female Speaker] And what happens if you have no HN? Can you express F? Does it get to the surface? [Speaker:Jennifer Platt] So this is what people did. So you can express F. It gets to the surface. And in certain artificial cell fusion assays that were done, like, you know, you can even induce fusion. I mean, you can induce F to fuse. You can induce F to trigger at high temperatures or even at 37. But there's a big difference between what you can get to happen in an artificial system and what really happens. But yeah, you can express F. It gets to the surface. It gets— you know, you can get it cleaved and so on. But it doesn't immediately go to postfusion complex. Without hn, it doesn't trigger. Well, I mean, it does. But it does. Well, you're already getting to the end of the story. But yeah, that— I know when I tell this now it sounds so obvious. Why did it take us decades to get to the truth? All right, so how is F-triggering kept in check prior to when you want it to be triggered? And so since HN activates F only upon receptor engagement, and we showed also that HN engagement has to continue until the onset of content mixing. So if you have receptor engagement, F gets triggered. It's not really this kind of spring-loaded mechanism that people call it. You can't then detach HN and just let F do its thing. HN has to stay stay there, stay interacting in this complex until there's really content mixing. And so what we showed biochemically in cell biological experiments where we co-express or individually express HN and F is that HN— the presence of HN prevents inappropriately timed activation of F, and the presence of HN retains F in the metastable state prior to receptor engagement. We showed that in cell expression experiments where we can modify the conditions and modify the availability of receptor and block receptor engagement. We were able to show that when HN and F are expressed on cells. But what we didn't know and what nobody knew was what is actually happening on the surface of the virus. How does this dance play out? And so plenty of structures of soluble HN and F had been done for parafluorid but also for other paramyxoviruses. But there was no HN-F complex. Nobody had been able to get an HNF complex. And nobody had been able, therefore, to look at what was going on on the surface of a virus. And so we set out to do this by cryo-EM and cryo-ET. And we developed a system, which I can tell you about in more detail, where we took authentic clinical viruses. And without purifying them or perturbing them in any way that messes up the glycoprotein delicate glycoprotein coat. We captured them on cryo-EM grids using antibodies strategy. We captured them gently on cryo-EM grids so that their glycoprotein coat is preserved, and we looked at them. And we also developed a way to make red blood cell fragments because red blood cells bear sort of authentic sialic acid-containing receptors. Not that the virus infects red blood cells, but red blood cells do contain from a human, you know, a panel of sialic acid-containing receptors. And we made fragments of that to interact with the viruses that are captured on the grids so that on the grids we can look at each of the steps of interaction. First, prior to receptor engagement, and then receptor engagement, activation of F, and complete fusion. And we can look at each of these steps first by cryo-EM and then by tomography. So what I'm going to focus on today is looking at this step right here when HN and F are presumably together on the surface of the virus in a complex and not yet engaged with receptor because we want to answer that question of how is that complex kept, how is F kept from triggering before the right time and place. And so going from this image here, my postdoc, Tara Marsink, who's done really all the work I'm going to show you now, did cryo-electron tomography on the complex of HN and F on the surface of a virus. So what you can see here is this is not interacting with the receptor. Prior to engagement. Here is F, and here are the heads of HN. You can't see any of the stock of HN. HN is— the stock is very flexible. I'm going to show you some of that in a minute. But you can see a little bit of the stock. This is at just under 10, so just sub-nanometer resolution with cryo-ET. And so is this clear to everyone? Here is this F layer here, and the HN layer here and that's, well, shown here. And so F is in its prefusion conformation and HN is sort of towering over it. So looking at that again here when it stops.

Is it interacting with all three pink monomers? Yep.

I'll show you that. We'll zoom in in a minute. There's a lot more of this I'll show you. But just if we decrease the resolution for a minute, you could see that you can just decrease the resolution so that you can see where the membrane is. You can see we don't have the stocks going all the way down to the membrane. So using AlphaFold Tera put together the whole picture. If you just use AlphaFold to predict predict where the HN heads will be, they'll be down here. You have to actually pull it up and put it where it belongs, and then you can get sort of a complete picture of the whole complex. And— So each dimer has a sialic acid binding site, but it looks like only one of them is interacting. Yeah, I'm going to show you that. Yeah, yeah, exactly. So I'm going to— here's what I'm going to show you. I'm going to show you that the HN is a dimer. And the dimer is in complex with the F trimer on the surface of the virus. But there's this loop on HN that reaches down to interact with the F apex and that each HN head in the dimer points in a different direction. One is pointing down to push on F and the other is pointing up to grab receptor. I'm going to show you that in more detail, but that's just kind of where we're going. So this is also, this is an overview of the dimer organization on the viral surface and I put this this and for Erica because they really are dimers. They are— you can see that— I mean, this is pretty— this is low resolution still, obviously, but you can see that there's probably some kind of lattice that we're working to understand, but that they are not tetramers. There are no tetramers. They're really— they're all dimers. And this is like flipped to the side, and you can see the HN layer. And what you can kind of see, the globs, blobs here, are the F's. And then you see the membrane. So getting back to this, zooming in on this region where there's a loop that's reaching down from the globular head of HN to the middle of where the trimers meet, and this is with the cryo-EM density and taking away the density, you can see this loop. And turning it around, you can see it from the side. And it's a loop. That is directly nestling into this pocket at the apex of F. And if you look at the whole complex again, you could see where this— I'm sorry, I just keep showing this picture so many times because I love it. But you could see really where that is in this whole complex. Peter Kwong had done a structure of a neutralizing antibody. Here interacting with F and published this. And here's the FAB, his Peter Kwong structure of FAB interacting with F, and it's a very neutralizing antibody. And remarkably, what we saw is that it interacts with F, Paraflu-F, in the exact same overlapping domain as HN interacts with F. This antibody works, we showed in an experiment where we look at F stability over time by stabilizing F. So with the antibody, with temperature, you can't trigger F as well as you can without it. This antibody, the Peter Kwong antibody that has crystallized and interacting with that exact same site, is stabilizing that prefusion F. And we think that the HN loop interaction is doing the exact same thing.

[INAUDIBLE] is not in a complex with HN? Otherwise, how does the antibody bind?

Well, the antibody— I wasn't going to show this, but the antibody displaces HN. The antibody has a higher— the antibody, if you—

So there has to be an equilibrium between on and off.

The antibody is— well, or the antibody is more avid for that site and is able to displace displace HN or is able to— we can't see it physically displacing but is able to bind and HN is kicked off. So here's the F antibody complex and the F antigen. HN complex remarkably overlapping. Now we had shown, just so we said, okay, there's this loop that seems really important. It seems to be a really potentially important key to how F is stable and to how F gets triggered at the right time and place. But we had shown in our last decades of work that when you express HN and F of various paramyxoviruses, you could swap heads. So you can take, you You can take a measles— you can take a parainfluenza head and put it on a measles stalk and get it to activate measles F. You can take a parainfluenza head and put it on a Nipah stalk and get it to activate— on a Nipah G stalk and get it to activate Nipah F. You can swap those heads. You could really swap them any which— almost any which way. All that you need is— all you need to activate an F is the right stalk of the right paramyxovirus. Is that clear? So we said, how could that be? How could that loop be so important? The sequences in that region are completely different. And if this is true, then how could we swap heads and it doesn't matter? Well, the sequences are completely different, but that loop is entirely conserved structurally despite the fact that the sequences are all different. And this is an overlay of all of them. Here they are individually. You can see parainfluenza, measles, Nipah, Hendra, parainfluenza, NDV. They all have that loop.

I guess you're proposing it makes backbone interactions then. What? I guess you're proposing it makes backbone interactions then because there's a lot of diversity of charge.

We have no idea yet. It's incredible though how the structure is completely conserved and you can use all those heads to complement all the stocks. And you can— yeah, so we're— it's obvious what we're trying to do now. Have you tried putting in different amino acids into this loop? No, this is brand new. This is not even published yet. I mean, this is— we're just starting it. And we need— we want to talk about to people who think about this stuff, like how we should do that in smart ways. Because we're just starting. Yeah, we have— we are making a bunch of changes and trying to understand exactly how this is working. I'll show you something else now.— Yes, exactly. It's matched. And yes, we've looked at whether or not they're like cooperative of evolution or anything like that, and we— it is matched. Yes. So that's a good question too. And we have really interesting mutations at that site in AFT that we got by selective pressure of growth with the neutralizing antibody because we figured we'll get changes in that region that'll be the same region that HN binds. And so we have very interesting changes in F that we're looking at because that pocket, yeah. So back to the thing that we, question that we were answering before. So the HN globular heads are pointing in two directions. So the head is, one head is pointing, sort of to grapple F, and the other head is 87 degrees rotated to engage sialic acid receptor. And when you turn it this way, you can see that that site 1, that sialic acid binding site, is rotated from the site that is facing up. So do we know that it's actually the distal to F head that is engaging receptor while the other head is engaging F? No. I mean, it's possible that when they get to receptor, that F grappling head, you know, flips up and— we don't know. But it's tempting to think that, you know, these two functions are being done by the two heads. And we're going to figure that out. And here's a molecular dynamics view that includes the stock, the taramate of the complex before receptor engagement. Okay, so just to wrap up this part, hn's capping F on the surface of authentic virions. And again, this is all with, like, right out of a supernatant fluid of clinical virus, preventing F from activating at the wrong time. Small molecules that perturb that hn dimer interfere with this mechanism, and you can see where we're going with this. Is there a rotation at that dimer? Is this— we're excited by the sort of convergence of what we're seeing in the structure and what we've known from the premature triggering molecules. And of course the site of HNF interaction at the F apex suggests a really exciting target for antiviral intervention if we can get at it. Just a little bit about the next step, which is— we're just starting where F is getting triggered by HN. And we could see these structures stretching across from the virus to the target red blood cell membranes that I mentioned that have sialic acid-containing molecules. Again, this is going to be much lower resolution, but here we are with that virus interacting with the target red cell membrane. You see the structures reaching across. They have the right height. They're intermediate F extensions. And here's a low-resolution image of the extended— by cryo-ET of the extended F reaching into the target membrane. And all I want to say about this is while it's low resolution and we're working on this, it's— what's really exciting is that we can also see HN that is still there that has moved down off its perch on the top of F. And this is where we are, what we're working on really heavily right now.

I mean, that's really interesting because first, so that you might need all three to make sufficient interaction site, and it seems to still be associated with one. You see some stalk, so it's not a stalk-stalk mediated interaction.

Right. Right.

Right.

And a while ago we'd already measured like the different heights, and we knew that there was some height changes going on, and the HN was like changing, but now hopefully we're going to be able to exactly see what's going on. Yeah. So this step, this step of extended intermediate followed by part— followed by folding has been where we've sort of done most of our antiviral work, blocking this late conformational step where we— where F needs to fold over on itself and where the heptad repeat domains, either and meet and the membrane merger. And we have done that for many different viruses, paramyxoviruses, measles, and non-paramyxoviruses, and developed small inhibitory peptides that interfere with that folding, lipopeptides. And we latest did it for SARS-CoV-2. SARS-CoV-2 is a spike protein that works very much like the the fusion protein of paramyxoviruses except the receptor binding portion leaves early in the process so that S1 binds ACE2 and leaves. So unlike paramyxoviruses where you have the receptor binding protein stabilizing the fusion protein, once S2 is left alone, it's all by itself. So it's actually very much floppier and more, more, has more range of motion in its intermediate states. Than a paramyxovirus F. But we did do the intermediate structures of those spike intermediates, and we did use a lipopeptide in red to immobilize, to interfere with that fusion and to prevent the zippering of the spike protein at the very final step and prevent fusion and actually inhibit prevent progression of infection and actually completely inhibit transmission of SARS in animals. So this is a strategy, this sort of preventing the final step of fusion is a strategy that's pretty broad. And we've used fusion inhibitory lipopeptides in vivo for, you know, Nipah, parainfluenzals, these are non-paramyxoviruses, and then cell culture for others. And it is just a general strategy that we're developing into a platform inhibiting that final step of membrane fusion. So basically we, as Erica said better than I can at the beginning, we look at these different steps and we try to elucidate the different steps in entry with the idea that they also offer targets for intervention. We hope to replace these sort of antiquated schematics with what's actually going on. And the first step that we've done that for, that we are in the process of doing that for, is this prior to receptor engagement step where we can see really excitingly aspects of the complex that we think are going to really help us develop ways to interfere with its function. There are a lot of people that contribute to this work. I didn't show— I just showed sort of a really focused piece today, but in our lab, Mateo Parado, Tara Marcinksa, postdoc, has done all the structural stuff. Peptide work, we collaborate with Sam Gelman at University of Wisconsin. Animal models, Stefan Niewiszka at OSU. Nir Bental is the computational who is at Tel Aviv University. We collaborate with Alex Granger for all our genomics. And Jose Malero, who I don't know if any of you knew, but he developed an antibody, the post-fusion antibody that allows us to— that does, in a lot of our work, allows us to see when F has been triggered and to prove how stable F is and what what affects its stability either in itself or by each end. And he died early, but we always think of him every day as we study F. And I hope that I didn't rush through that. I hope we have time for questions now.

So all the crystal and cryo-EM structures of the HN, the green protein, the tetramers. Why is it always a tetramer even without crystal packing?

And then there's a mix of downs and ups. And well, you see the surface of a virus. I mean, I've never done crystallography. I don't, I don't know. But we're looking— there are so many. First of all, I mean, there's just so many artifacts even when we, even if we pellet this virus, for example, there's just so many artifacts that glycoproteins are a mess. I mean, I just, I don't know. I think all we know is that we are absolutely not seeing tetramers. Why do you think they were?

I don't know. Maybe two of them interact on the surface of the virus.

Maybe it's satisfying something that— Well, we definitely think that the, I mean, we do have evidence that there is a lattice on the surface of virus.

But maybe they do this in the absence of an F, right? People have observed HN.

People have observed F. But they've never seen them together. Yeah. Who cleaves F? Which enzyme cleaves F? Well, there's a dogma about that, which is that furin cleaves it during transit through Golgi. We're actually going to change that soon, we think, but I can't— because that's what happens in tissue— in lab-adapted viruses in tissue culture cells. Any other questions for Anne?

Are you going to end up making any stock mutations or—

We have— you mean HN stock mutations? We have HN stock mutations that we've studied, and they've really told us a lot about the role of the stock in triggering. The stock is key. You have to have the stock paired with the fusion protein to get triggering. And how are we going to figure out how that works if we can't see the stock? That's the next— we need like— yeah, there's so much that we we could do, but stock is very interesting and we have ideas about what it's doing. Clearly the whole molecule is shortening. What's stock doing? Like, we got to figure that out, but—

[INAUDIBLE] —what kind of—

is tetramer active, dimer active, or monomer active?

Well, we don't think there are tetramers. So, yeah, dimer. So— If it is expressed as a monomer, is it active, like cleaving the sialic acid? Is a monomer of HN active at cleaving sialic acid? I don't know. I don't know how you would— Yeah, I don't know that. I'm not sure how you would— right, you mean binding, binding, not neuraminidase. Yeah, because, you know, HN does have Right, and RNAse II. I don't know where you— I don't— yeah, I'm not sure.

So in terms of this conformational change in F, if you just crystallize it as an independent individual protein, is it in this sort of mushroom shape naturally, or is it in the extended shape naturally?

No one's ever captured the extended intermediate. You can either capture the prefusion, so the— before any receptor engagement state, or the postfusion when it's already extended and flipped back and folded on itself. So we're the first to get images of it, glimpses of it in its extended intermediate, which is very transient. Nobody's— there's no crystal structures of that.

So you think HN is somehow— suppressing its ability to form this extended structure and you change— when HM binds sialic acid, it changes its conformation and that just releases F?

No, we're not saying it releases F. I mean, we definitely— we do think there's a protection prior to the receptor binding. Receptor engagement. We do know there's an active triggering process where HN activates F, and we do know it's not enough to just release it because it has to be there all the way through until complete membrane merger. So if you get F triggered and inserted and then you disengage HN from its receptor, fusion halts. Like, it's not enough. It's not like just spring-loaded. So we think there's a— we think HN is doing many different things. It's not just that it gets out of the way and NF can go do its thing. No. We think it's— we know that it's doing more. We just don't know how yet. We think we know a little bit about how it's protecting it. And we know parts of the molecule that are necessary for the active triggering stuff. We know that dimer interface residues are necessary. Necessary for the triggering step. We know specific residues in the stalk that are necessary for the triggering step, but we don't know how. So for me, structures are like how. Structures are really showing us how. They're not just showing us what things look like. They're like, how on earth is that happening? And like— Erica looks very perturbed. The equine, the HEPA vaccine, is it the hemagglutinin or the fusion protein? CRISPR editors. Oh, I think it's fusion. I'm not sure though. I mean, it's interesting. It seems like if it's under this lawn, like maybe the HN should be the main component. Well, the neutralizing— well, like the Peter Kwong neutralizing antibody, like when you have an antibody, you can get a— get HN off and neutralize F. So I mean—

[Speaker:AUDIENCE] But is it like different layers of fluid motion? It's interesting that the way I was thinking about the more hidden one down here instead of the obvious one at the top. Any other questions?

[Speaker:AUDIENCE] Well, that's interesting you showed here.—

that can control movement. So have you ever synthesized that group and tried the affinity of this F, or even this short peptide can provoke motion? No, we really— this is really new, so we haven't done— that'd be a great idea. We want to— design lots of different things to get into that area and try to block. That would be one of them for sure.

Thanks. For measles vaccination, is it known whether the F or HN are the targets of the more protective antibodies?

You mean for the live, for the vaccine that we're using now? Yeah, I don't know. I should know, but I don't know. I think it's not known. Oh really? Yeah. Maybe that's why I don't know. Not known, right? There's a lot of epitope mapping for mouse antibodies and you get primarily antibodies against H and then like a couple against F.

And if you do adoptive transfer of those antibodies, um, I don't know about monoclonal, but passive transfer of polyclonal antibodies, no, no, I was saying if you have a monoclonal specific for HN versus F, you could sort out which one would act better.

I don't know how much work has been done. Me neither. I think it's such a widely known— yeah. Oh, questions from—

in the chat? Oh.

How do you look in the chat? I guess I can stop sharing. Yeah. 7.

Oh, oh, oh, what am I doing? Okay. Hearing, no, no, no, no, no, working on a success.

All right, taken care of. Zero questions? All right, if no other questions, let's thank Annie.