VIDEO: Nicolas Locker on Friends or Foes? The Many Routes Caliciviruses Use to Manipulate RNA Granules

Nicolas Locker (00:03:34):
If I can have the next slide please. Thank you. Really, the fundamental questions that we try and address in my lab is the role of viral RNA in hijacking the cellular translation machinery, and equally how host mRNAs are regulated by virus signaling and stress pathways during infection because if you think about it, a virus is a parasite. It has to use the cellular resources in order to replicate, and there is always that conundrum to resolve. How can we balance the production of viral proteins with the adverse effects of producing cellular protein for the viruses? Right?

Nicolas Locker (00:04:26):
On the one end, you need to make viral proteins. On the other hand, you need to make cellular proteins. And so in a nutshell, we try and address to answer two questions in the lab. How do viruses translate their genome, and I’m not going to discuss about that today, and what is the impact of viruses on host mRNA translation?

Nicolas Locker (00:04:45):
It’s by that interesting mRNA translation that we got our finger caught in the stress granule cog a few years ago because obviously as soon as you start to look at stress responses and translational control with awareness of the impact of biocondensate on these processes, we very quickly had to jump in the stress granule regulation by viruses. Next.

Nicolas Locker (00:05:19):
I guess for this audience, I don’t really need to emphasize the role of condensation in cellular processes, so next slide, please. But, I guess the only thing I want to point is that there’s really a… I would say in the cell biology field, an explosion in the appreciation that controlling the localization and function of biomolecules is really fundamental for many cellular processes, and stress granules are really a paradigm for understanding how membraneless organelle or biocondensate work during stress. Next. Because the impact on signaling, translation, and metabolism, and with the hypothesis that heterogeneous stress granules can basically drive poor survival or poor death function in the cells. Next slide, please. So here actually that’s going to be easier if we click a couple of times to get rid of the animation, please. Yeah. Perfect. Oops. Go back one. Perfect. Thank you, Ryan.

Nicolas Locker (00:06:39):
So in response to viral sensing, it’s well-appreciated that the activity of specific translation factor can be controlled either via their proteolytic cleavage, by viral protease, or through their phosphorylation. This results in the accumulation of stored initiation complexes that are recognized by aggregation-prone RNA-binding protein, (G3BP1, Caprine 1, TIA1, to name a few), and all of these complexes kind of collapse and condense into stress granules.

Nicolas Locker (00:07:11):
I think for virologists what’s interesting is that it’s been shown, and there has been the hypothesis, that stress granules can contain sensor and effector of the antiviral response, RIG-I or PKR to name a few, and therefore it has been proposed that they could act as a platform for antiviral activity. Of course, they’re also associated to adaptation to the tumor microenvironment, and persistent stress granules are associated with neurodegeneration. Next slide, please. Yeah. Again, here we can click the animation. Perfect. Thank you.

Nicolas Locker (00:07:51):
I think one of the key points to appreciate for this talk is that in response to… There is really a virus-specific and a cell-specific phenotype of granular bodies, if you will, where it’s been shown already that during infection few different types of foci can assemble. On the one end, we can have antiviral stress granule that assemble early during infection before the viral countermeasures kick in, or that can assemble in response to mutated viruses that are enabled to suppress stress granules. They are associated with increased interferon production and impaired replication. They condense and activate innate immune sensors and effectors. They’re also antiviral protein.

Nicolas Locker (00:08:43):
And at the same time we have other bodies such as RNase-L bodies that have been shown to form in response to poly (I:C) treatment and RNase L activation. They form independently from elF2 alpha phosphorylation and translational inhibition, and very importantly they do not require the key scaffolding protein, G3BP1, for their condensation. They are more associated with mRNA decay than translational storing and dock with P-bodies later in infection to stimulate the processing and degradation of cellular RNA. Next slide, please.

Nicolas Locker (00:09:26):
I’ve summarized here the kind of different properties of these bodies, and I’ll actually come back to that kind of snapshot summary in the last part of my talk when I discuss the novel SG-like foci that we recently described. Next, please.

Nicolas Locker (00:09:44):
Of course, because of this antiviral effect, viruses have evolved countermeasures in order to be able to replicate, and here I’m definitely not going to go into all the details. That’s a figure I extracted from a review that we have coming up in, I guess, a month that kind of summarizes the many ways that viruses can disassemble or prevent the assembly of stress granules because, of course, these stress granules, whether they have antiviral activity, per se, or simply are associated with translational inhibition, that is deleterious for viral translation and viral replication. Next slide, please.

Nicolas Locker (00:10:28):
So over the past, I would say, maybe going back to the RNA Society meeting in 2014, so eight years, my lab has been really trying to understand using Dengue virus, or flaviviruses, norovirus, how viral infection regulates the mechanisms by which viral infection regulates stress granule formation. But I guess one of the things that we’ve been doing also recently is to challenge the view that stress granules only contain RNA and proteins.

Nicolas Locker (00:11:06):
I will point you to the really cool study that we published a couple of… I guess about a year ago now with colleagues at the University of Gothenburg in Sweden where we used amperometric measurement and fancy electrochemical measurements to reveal that reactive oxygen species, so reactive molecule, are stored within stress granules. So they don’t just contain RNA and protein. They can also contain, let’s say, potentially secondary messenger, and that could explain how they impact on signaling.

Nicolas Locker (00:11:48):
It’s always quite nice when you see editor highlighting this paper as very important paper. That’s not me writing that. That’s the editor, so that was really a cool addition to our stress granule success story. So next slide, please. Yep. Four clicks. Perfect. Thank you.

Nicolas Locker (00:12:14):
The take-home message from today’s talk, and if you’ve got something cooking, if you’re busy with other meetings, these are the take-home messages that you should take away. What I’m going to do is show you that norovirus infection triggers metabolic stress rather than canonical antiviral sensing. This contributes to inhibiting stress granule assembly by repurposing G3BP1 and kind of rewiring the G3BP1 interactome. I will show you that another calicivirus impairs stress granule formation in a different way by cleaving G3BP1, and finally I will describe a novel SG-like foci that we identify during infection that we named paracrine granules. Next slide, please.

Nicolas Locker (00:13:06):
So with that, I will just have one slide about the virus to introduce it. So our models are norovirus which are members of the calicivirus family. I guess in a way I don’t need to elaborate too much on the symptoms that are summarized on the bottom on the right-hand side. The bottom line is norovirus, you really don’t want it. It’s very infectious. It’s responsible for most of the gastroenteritis cases worldwide. It transmits very easily, fecal-oral transmission.

Nicolas Locker (00:13:44):
It spread very rapidly in enclosed settings, and although the symptoms only last for two to three days, and if you have kids that attend school or nursery, I guess in the US like in the UK, the guidelines is, “Okay. If your kid is puking, he has to be off the nursery for 48 hours,” but in fact, they will be shedding viral particle for two weeks which kind of explains why as soon as you have one child sick or somebody sick in an enclosed setting, in an enclosed facility, so like hospital or a care home, there is an explosion of cases because you can’t basically keep people or prevent people from going to work for a couple of weeks.

Nicolas Locker (00:14:29):
Because of that absence of staff from work, it really costs the healthcare system in the UK 100 millions of pounds every year, and we often read about outbreaks in restaurants, in hospital, even in kind of famous athletic events, so the Olympic games in PyeongChang. I think the entire Finnish skiing delegation was wiped out with Noro. At the London 2017 Athletics World Championship, we had the guy who in the end won the silver on the 200 m, went to run by himself because he had norovirus, so to avoid contamination to others. Really, a bit of an impact on the population. Next slide, please.

Nicolas Locker (00:15:25):
So in previous work, what we had done was to demonstrate that Norovirus has the ability to impact the activity of specific translation factor, either by cleaving those factors via viral protease or activating their phosphorylation through the activation of intracellular signaling pathways such as p38. Next slide. So the question we then asked was what is the global impact of his regulation on the translation during infection? Next.

Nicolas Locker (00:15:59):
What Michele in my lab did was to use ribopuromycylation assays that allow us to measure translation efficacy at the single-cell level to basically demonstrate that norovirus induces a gradual shutoff during infection, so if we have, I guess, three clicks here. Yeah.

Nicolas Locker (00:16:22):
What Michele was able to demonstrate was that if you look for cells that… I’m sorry. I haven’t got a pointer to show you that, but if you look at cells that are positive at 12 hours for the NS3 viral protein, these cells don’t light up all with puromycin. Right? If you look at the merge signal, the cells that are yellow don’t have any magenta really displaying that strong shutoff of translation. We know from a lot of studies that during stress and infection, it’s the phosphorylation of elF2alpha that drives the inhibition of translation. Next, please. Next.

Nicolas Locker (00:17:06):
Of course, we looked at elF2alpha phosphorylation during infection, and we were able to observe that elF2alpha is phosphorylated during MNV infection while an inactivated virus doesn’t trigger it. Next, please. We then ask, “Okay. Is this phosphorylation really what drives the shutoff,” and to answer that question we used mouse embryonic fibroblasts that are either wild type or that have the phosphorylation site mutated to an alanine, and what we could show if you focus on the quantification on the right-hand side is that basically the translational shutoff occurs whether elF2alpha can be phosphorylated or not, which told us that basically the stress-sensing via elF2alpha is uncoupled from translational stalling. Next.

Nicolas Locker (00:18:05):
So to cut a long story short, we then try and identify the elF2alpha kinase responsible for this phosphorylation, and we anticipated that the usual candidates, PKR or PERK that are associated with viral protein overload or viral RNA sensing to be responsible for elF2alpha phosphorylation, but what we found if we can click forward once was that the inhibitor of GCN2, A92, was able to revert elF2alpha phosphorylation (so that’s the top blot on the right), and that using MEF knockouts for GCN2 (so that’s at bottom Western blot on the right), we could see that in the absence of GCN2 we did not observe this phosphorylation of elF2alpha during infection really confirming that it’s GCN2, the metabolic sensor, that drove elF2alpha phosphorylation during infection. Next slide, please, and next.

Nicolas Locker (00:19:16):
So that kind of lead us to suggest that maybe infection could trigger a metabolic response rather than normal viral sensing, and this is what we set out to investigate. Next.

Nicolas Locker (00:19:33):
So the first thing we did was to measure total amino acid levels in infected cells, and we couldn’t see any difference. Next slide. However, when we quantified using LCMS, the levels of individual amino acids, what we could see was that infection would result in the specific depletion of a particular subset of amino acid. If you think about it, that makes sense because viruses have a very small genome. This Norovirus only makes eight proteins. The distribution of amino acids within those proteins is biased, so basically what you’re seeing here is a depletion that corresponds to the virus just eating up preferentially a subset of amino acids during infection. Okay. That could explain the metabolic stress. So next slide, please.

Nicolas Locker (00:20:29):
So we know from previous study that amino acid starvation is associated with signaling via GCN2 and signaling via JNK2, and by using antibody arrays we could detect JNK2 activation and ATF2 phosphorylation. Next. The up-regulation of their downstream transcript really convincing us that we could observe a starvation associated with immunosuppression as well since we observe a really high increase in the immunosuppressive cytokine IL6. Next slide, please.

Nicolas Locker (00:21:13):
So we used RNA sequencing to characterize a little bit better the genetic reprogramming, and we were able to really observe the transcriptomic signature associated with amino acid starvation, with the up-regulation of many ISR targets, and no up-regulation of the targets that are usually associated with pro-inflammatory signaling and antiviral response, so really something that fitted amino acid starvation. Next slide, please.

Nicolas Locker (00:21:51):
Just to pick on a specific example, GDF15, which is one of these ATF3 targets up-regulated, when we neutralized GDF15 with antibodies in the supernatant or during infection, and these are the magenta lines, we could see a rescue or an increase in survival during infection. Next please. This was associated with an increase in the production of pro-inflammatory cytokines, so neutralizing this amnio acid stress response would induce inflammation infected cell, which basically tells us that the role of this amino acid stress is really to maintain the cell in an immunosuppressed state so that the virus can replicate. Next slide, please.

Nicolas Locker (00:22:50):
So we can come up with a model in which infection with noro induces metabolic stress that activates GCN2 and JNK. We have a transcriptional reprograming that’s associated with the immunomediatory response and amino acid starvation to create an immunosuppressive environment that favors replication. Next slide, please.

Nicolas Locker (00:23:16):
What I’ve told you so far is that norovirus infections remodel host translation, that elF2alpha phosphorylation is uncoupled from the shutoff, and that metabolic stress sensing is important to favor viral replication.

Nicolas Locker (00:23:30):
So what’s the connection to biocondensate, you’ll ask me. And obviously metabolic stress is a stress, so we wondered what was the impact of that stress and of the shutoff of the SMB of stress granule during infection. Next slide, please. Next.

Nicolas Locker (00:23:52):
So we used… I’m assuming that a lot of you are familiar with these tools and techniques, so we used cells expressing a GFP-tagged G3BP1 to monitor stress granule assembly showing that G3BP1 in our cell models, so these are microglial cell that are infected by norovirus. G3BP1 locates to stress granule following arsenite stress, and these granules are dissolved, as expected, by cycloheximide treatment. Next slide, please. Okay. Yeah. Now, if we look at the distribution of G3BP1 during infection, we can see that it accumulates in cytoplasmic foci, but it also seemed to co-localize with the viral protein NS3. Furthermore, when we treat the cell with cycloheximide, this co-localization is maintained. Next slide, please. Which kind of told us that potentially here we were looking at virus-specific granules. Next slide.

Nicolas Locker (00:25:03):
So we went a little bit deeper, and here what I’m showing you is that we were able to demonstrate that during infection, and these cells are very round in shape, and it may look like how the co-localization occurs in the nucleus, but actually that co-localization that looks like a little donut sits on top of the nuclei, so these condensate are really cytoplasmic.

Nicolas Locker (00:25:33):
We were able to show that basically G3BP1 relocates to replication complexes during infection, and this is how MNV basically evades stress granule formation. When we looked at the interaction of G3BP1 with other potential stress granule proteins, we could see that we were still able to IP some stress granule markers such as Caprin1, but we were also IP-ing more strongly USP10, a negative regulator of stress granule assembly, and obviously a big marker for stress granule condensation initiation factors. Here we could see that eIF3 was excreted from this granule really confirming that we are looking at bodies that are definitely not stress granules. Next slide, please.

Nicolas Locker (00:26:26):
We then went on, and I’m not showing you the proteomic data here, but we went on to establish the composition of these granules and showed that they are really completely different from canonical stress granule induced by arsenite. They have their own components that only co-localize in the MNV granules as RAB7b in panel C, and what’s more important is that when we started to knock down, for example, SF3B1 or RPL7 which selectively localized to the replication complexes, we strongly impaired viral replication. So the assembly of these viral-specific bodies are important for viral replication. Next slide, please.

Nicolas Locker (00:27:20):
Is that conserved across related calicivirus? Next, please. That’s a question that a really talented PhD student in my lab, Majid Humoud, undertook where he looked at the assembly of stress granule in feline cells infected with the feline version of norovirus, FCV, and he could demonstrate that during FCV infection we observe elF2alpha phosphorylation, right-hand side on the top, but if we click a couple of times, so next, next. Yeah. Once more, please. Thank you.

Nicolas Locker (00:28:03):
What Majid could observe was that during infection, A, G3BP1 would remain cytoplasmic, and that if we were to prime the cell with sodium arsenite after infection, these cells would be unable to assemble stress granules really showing that FCV infection disable the assembly of stress granule, and that could be reverted by inactivating FCV with UV radiation. Next slide, please.

Nicolas Locker (00:28:35):
Again, I’m trying to cut a long story short here, but these viruses expressed a protease called NS6, and we were able to show that during infection G3BP1 is cleaved. Next please. Oh. I’ve lost an animation here. Oh. Yeah. Perfect. Cheers. By over-expressing the Feline Calicivirus protease, we were able to show that the protease is indeed responsible, so that’s panel B, for the cleavage of G3BP1. And interestingly MNV which inhibits stress granule by repurposing G3BP1 in replication complexes doesn’t have to cleave G3BP1 because it’s already disabling stress granule assembly, so the protease hasn’t evolved that targeting capacity for G3BP1. Next slide, please.

Nicolas Locker (00:29:33):
Using mutagenesis, we were able to map the cleavage site at position 405, and you can see on the merged bottom panel that basically mutating this cleavage site to an alanine that is not cleavable anymore by the protease rescued the formation of stress granule.

Nicolas Locker (00:29:58):
Why do we care? How different is that from other studies that have been done on viral entero protease? Next slide, please. Here I’m displaying the cleavage site for poliovirus and many other enteroviruses and the cleavage site that we have just identified, and what you can see is that interestingly what FCV is doing is cleaving only the RGG domain of G3BP1, but not the RRM and the RGG domain by poliovirus.

Nicolas Locker (00:30:29):
In the context of recent papers from [inaudible 00:30:34] and then [inaudible 00:30:35], it’s really interesting because what FCV is doing is being super-targeted and only cleaving the G3BP1 domain that’s important for binding the small ribosomal subunit during the condensation process. Instead of cleaving all the RNA interaction domain like other enteroviruses, it really targets the one domain that’s fundamental for G3BP1 condensation. Next slide, please.

Nicolas Locker (00:31:07):
Here we can come up with a model in which upon translational stalling, the two different or the two related viruses are using different strategies in order to block stress granule formation. On the one end, MNV hijacks G3BP1 while FCV cleaves G3BP1. And in parallel to that to maintain viral propagation, the virus induces a metabolic and amino acid stress response that promotes replication and propagation. Next slide, please.

Nicolas Locker (00:31:45):
So I will come now to the final part of our talk which, again, came from interesting observation from this PhD student, Majid, where… Next slide, please. Where when he was quantifying stress granule assembly in infected cells, he was always able… So the infected cells are labeled in purple. He was always able to detect isolated cells in there with stress granules to a level that was above the significance level.

Nicolas Locker (00:32:21):
These cells were always uninfected, so we didn’t really pay too much attention to them because we were so focused on the protease cleavage and mechanism, but we came back to this data a little bit later, and we started to wonder. Could there be some signaling mechanism going on from infected cell towards uninfected cell, so towards bystander, in order to induce a stress response, some sort of call of duty or alarm bell ringing to tell an infected cell, “Okay. There’s a virus coming. Shut down everything to prevent replication.” Next, please.

Nicolas Locker (00:33:07):
This kind of lead us to hypothesize that perhaps there was some paracrine signaling going on, and so to test this hypothesis what we did was to take cells and either mock-treat them, so I know it’s a word that my colleagues often struggle with, so we virologists tend to call mock the noninfected control, so we use either mock-infected the cells or infected cells with FCV. We collected the cell supernatant. We removed the viral particles to end up with what we call the VFS or virus-free supernatant. And that supernatant is free of viral proteins, it’s free of viral RNA, and if you stick that supernatant on live cell for three days to a week, you do not recover any viral particle. So it’s whatever signaling messenger are produced during infection minus the virus. Next slide, please.

Nicolas Locker (00:34:14):
What we did was to take that supernatant and stick it onto cells that had never seen a virus and look for stress granule induction, and lo and behold, in the middle panel what you can see is that typical stress granule markers such as G3BP1, FXR1, and UBAP2L re-localized into cytoplasmic foci when cells are treated with virus-free supernatant. Next slide, please.

Nicolas Locker (00:34:45):
When we quantified the characteristic of this SG-like foci, let’s call them like that for now, we could see that they were basically present in higher numbers as compared to the number of stress granules when cells are stimulated with arsenite, and the size of the granule is smaller, so we have more of them, and they’re smaller, so they don’t really share the properties of normal stress… Of, let’s say, type 1 stress granule. Next slide.

Nicolas Locker (00:35:21):
We then wanted to see whether they were sensitive to cycloheximide treatment, and you can see in the middle panel that basically the foci still assemble, are still assembling when the cells are treated with cycloheximide, so they don’t seem to rely on trafficking of mRNA away from proteosomes into SG-like foci, which again is a property that is different from stress granule but is common with RNase L body. Next slide, please.

Nicolas Locker (00:36:00):
We then try and address their kinetics using recovery assays in which the cells were stressed with arsenite or stimulated with VFS for 45 minutes and then left to recover after washing out the stressor. Next slide, please. Here, again, a fundamental difference is that while it takes about an hour-and-a-half for arsenite-induced granules to disassemble, our SG-like foci dissociate much more quickly, and they can be gone in some cases by 10 minutes, but it’s between 10 and 30 minutes on average. Next, please.

Nicolas Locker (00:36:47):
Again, showing us that they are… If we go back, that their kinetics is completely different from SGs and from RNase L body. Next slide, please. Obviously, a fundamental property of stress granule: they depend on G3BP1. So if you knock down G3BP1 in cells, that’s the bottom panels on the right, you use lose the ability to assemble stress granule when cells are stimulated by arsenite, whereas when cells were treated by virus-free supernatant, whether G3BP1 is present or not doesn’t affect the formation of these SG-like foci.

Nicolas Locker (00:37:38):
By this point, we were really convinced that these are not stress granules and decided to name those bodies paracrine granules of PGs. Next slide, please. Again, you can see here a fundamental difference from stress granule. They don’t depend on G3BP1, but that’s a property shared with RNase L body. Next slide, please.

Nicolas Locker (00:38:04):
We decided to establish their composition using proteomics and RNA seq analysis, and I’m fortunate to have a longstanding collaboration with the Parker lab in Boulder, Colorado. I’ve spent quite a few weeks enjoying Boulder treats, local brews, but also learning how to purify stress granules using their immunoprecipitation techniques, and what you can see here are individual stress granules that are tethered onto dynabeads that are the suitable for RNA seq and proteomic analysis. Next slide, please.

Nicolas Locker (00:38:48):
I guess I’m not going to go into all the different pathways and all the details, but the bottom line for me here when we looked at the composition of paracrine granules is that there is very little overlapping composition between paracrine granule and stress granule. However, importantly they are enriched in proteins with disordered domains. They are enriched with proteins that have RNA-binding domain or domain associated with nucleic acid binding. Interestingly, in the context of the paracrine signaling, the PG-specific proteins are associated with functional pathways that are linked to cell-to-cell communication, signaling, apoptosis, and senescence which could have its importance in that preparation and preventing of viral replication.

Nicolas Locker (00:39:49):
Importantly, in our proteomics we could not detect any translation factor, so again really reinforcing that point that they are disconnected from, or they are not assembled, as a result of translational shutoff. Next slide, please. Again, a fundamental different property from stress granule. Next slide.

Nicolas Locker (00:40:18):
We then looked at the properties of the RNAs that are triaged into paracrine granule, and here I would like you to focus on the graph that has some pink and blue which basically shows you that the mRNA that are triaged into paracrine granule are longer than the mRNAs that are triaged in stress granule. Next slide, please.

Nicolas Locker (00:40:48):
When we looked at the identity of the mRNA, we saw a really vast difference. Some overlap, but also some difference in panel E in the identify of the mRNAs that are triaged into paracrine granules. When we looked in F at the functional pathways that are enriched in the paracrine or stress granule, again we saw an overlap but some differences. However, when we looked at the top-10 pathways, functional pathways, that are associated with these RNAs, in stress granules or in paracrine granules, and that’s panel G, what we could see was that these are in fact the same functional pathway.

Nicolas Locker (00:41:33):
So what does it tell us? It tells us that although the identity of the mRNA are different, these granules condense to similar functional classes of mRNA. Okay. Next slide, please. We then searched for motifs enriched in these mRNAs, and of course we were able to identify motifs, but more interestingly when we search for RNA-binding protein that preferentially recognize these motifs, and that’s what you see on the right-hand side of the table. What we could observe was that several proteins highlighted in gray or in black that have the ability to bind those motifs are actually proteins that our proteomics identified in the paracrine granule proteome. What does it tell us? It tells us that these RNA-binding proteins could be the ones responsible for triaging mRNAs into paracrine granule. Next slide, please.

Nicolas Locker (00:42:45):
Finally, going back to translation and signaling, we wondered whether paracrine granule assembly would be associated with translational shutoff, and it is, as we saw a really strong reduction in puromycin incorporation. Next, and next, and next, again, please.

Nicolas Locker (00:43:11):
And then when we looked at signaling, what we could see was that monitoring the activation of various signaling pathway, we saw that basically stimulation with virus-free supernatant induced ERK phosphorylation and elF2alpha phosphorylation. However, when we used either genetic inhibition of elF2alpha phosphorylation (that’s the IF panel on the left), or pharmacological inhibition of ERK (that’s the panel on the right), we could see that virus-free supernatant would still result in the assembly of paracrine granules, which tells us that the ERK signaling or the elF2alpha phosphorylation are not responsible for the assembly of the paracrine granule but potentially this signaling event occur downstream of paracrine granule assembly. Next slide.

Nicolas Locker (00:44:13):
Does it matter for the virus? That’s the big question. To answer this question, we took live cells, stimulated them with virus-free supernatant, and then stuck some virus back, and measured viral titre. Next slide, please. What we could see was that virus-free supernatant induces a dose-dependent inhibition of viral replication, so paracrine granule assembly is associated with viral replication.

Nicolas Locker (00:44:49):
I have to say importantly that this is independent from interferon signaling. Okay? There is no interferon detected in the virus-free supernatant. Interferon in itself doesn’t induce stress granule or paracrine granule in the cells we are using, so this is really down to paracrine granule assembly. Next slide, please.

Nicolas Locker (00:45:15):
Again, a whole bunch of properties that are different from other bodies that form in response to infection. Next slide, please. So what I’ve shown you is that we identified a new type of stress granule-like condensate that assembled in response to infection. They don’t require polysome disassembly or G3BP1. They are more dynamic than SG. They are enriched for different proteins. The mRNAs that are triaged into paracrine granule are paracrine granule-specific, but they are enriched for similar functional pathways than stress granule, and importantly the assembly is associated with shutoff and antiviral activity.

Nicolas Locker (00:46:05):
What don’t we know? We don’t know whether this is conserved across other viral model, so we don’t know if other viruses induce paracrine granules. The big question your going to ask me, everybody asks me that, what is the nature of this paracrine mediator? I haven’t got an answer. I’ve got some clues that I’m happy to discuss, but no answers yet. Next slide, please.

Nicolas Locker (00:46:27):
That leaves me to propose a final model, so I guess we can click a few times to have all the model on the screen. Thank you. Basically, early in infection FCV induces the formation of stress granules that are then inhibited or disassembled via the cleavage of G3BP1 by the FCV protease. Next please. And then this infection is sensed by bystander cell that assemble paracrine granule that prevent further replication and prevent further spread of FCV during infection. Okay?

Nicolas Locker (00:47:19):
That leaves me in the next slide to summarize the key points from today’s talk. Norovirus infection triggers a metabolic stress and not canonical antiviral sensing, and that promotes replication. Next. It inhibits stress granule assembly by repurposing G3BP1 and rewiring the stress granule interactome during infection. Next. This mechanism is not conserved among very closely related viruses because Feline Calicivirus impairs SG formation by cleaving G3BP1. Next slide, please. Caliciviruses induce the formation of SG-like foci in bystander cell that are associated with antiviral activity and that we named paracrine granule. Next slide, please.

Nicolas Locker (00:48:18):
And so finally I simply want to acknowledge the members of my labs that were involved at various stages of this study. Michele, Valentina, Belinda, Kushboo, and Lucy, and the wonderful collaborations that we have with Ruggieri Group in Heidelberg, the Parker group in Colorado, the Goodfellow group at the University of Cambridge, UK, and Martijn Langereis at the University of Utrecht, and UKRI, the Biotechnology and Biological Science Research Council for basically funding my lab for the past few years and hopefully for continuing to fund these fun stories on biocondensate and viruses. I will leave it there, and I’d be happy to take any question you may have about the different stories that I’ve shared today with you. Thank you.

Balaji Olety (00:49:23):
Thank you so much, Professor Locker, for that fascinating talk. We already have a lot of questions. First, Diana.

Diana Mitrea (00:49:34):
Thank you for that talk. That was beautiful. So I was wondering if you could comment a little bit about the types of long non-coding RNAs you found in your PGs and how that differs from the stress granules.

Nicolas Locker (00:49:50):
Yeah. Sure. So I guess it’s in terms of noncoding RNA, I have to raise my hand and say that this is not something that we have followed up extensively at the moment, but we do observe noncoding RNA that have been previously associated with stress responses, and I guess interestingly we also observed long noncoding RNA that are typically associated with paraspeckle assembly in the nucleus.

Nicolas Locker (00:50:24):
I’m not sure whether you’re aware of those. You’re probably aware of those papers and studies that have shown that in type I stress granule assembly, so granules that are assembled in elF2alpha-dependent pathways, there is that whole hypothesis that once the stress is removed, and stress granules have disassembled, there is then a second wave of stress sensing, this time in the nucleus, through the assembly of paraspeckles on the long noncoding RNA NEAT1 which traps other long noncoding RNA and regulates splicing activity.

Nicolas Locker (00:51:12):
You could have that model that first you have stress granule assembly in the cytoplasm. That controls translation, and then paraspeckle assembly in the nucleus, and that controls splicing. The fact that we see some of these paraspeckle targets in paracrine granule kind of lead us to hypothesize whether they could also be involved in some sort of cytoplasmic nuclear crosstalk by preventing or maybe delaying the assembly of paraspeckle. This is something that we are pursuing at the comment. Whether it is the case, it’s really too early to comment on that, so I’m not truly answering your question, but I’m trying to give you clues as to in which direction we are going with the data.

Balaji Olety (00:52:12):
Thank you. The next question is from Dr. Xiao Tong. Dr. Xiao, could you possibly unmute and ask the question?

Xiao Tong (00:52:24):
My question is, given the difference you see between MNV and FCV, I’m just curious. Have you looked at human noroviruses?

Nicolas Locker (00:52:36):
Yeah. Yeah. So we looked at-

Xiao Tong (00:52:41):
-because there’s a medical need there.

Nicolas Locker (00:52:41):
Yeah. So we looked at human norovirus. I guess it’s something that I didn’t discuss, but the reason why we focused on murine and feline norovirus, and the reason why a lot of calicivirus people focus on these two animal model for human noro is because it is extremely difficult to propagate human norovirus. There is no cell culture system, and you have to use… There are specific, primary B-cells that you can use again with models, but nevertheless we looked at the conservation of the interaction between G3BP1, and some of its specific targets within the replication complex during human noro infection, and we saw that this was conserved.

Nicolas Locker (00:53:35):
One thing that I did not discuss, but in the interest time, is the fact that why is G3BP1 relocated to replication complex? It’s because G3BP1 has the ability to bind the viral RNA and to increase translation of the viral RNA. This property is conserved between murine and human norovirus. Okay? So this, if you want, rewiring of the SG interactome and repurposing of G3BP1 as a translational activator during infection is conserved between murine and human norovirus.

Xiao Tong (00:54:20):
Thank you.

Balaji Olety (00:54:24):
Alright. Thanks. The next question is from Aravinthkumar Jayabalan. Would you unmute yourself and ask the question, please?

Jill Bouchard (00:54:35):
They had to run to a different meeting. Do you mind reading it?

Balaji Olety (00:54:37):
Sure. The question is are these paracrine granules containing cells induced stress granules at a later point during infection?

Nicolas Locker (00:54:51):
Huh. That’s a good point. I would say, “No.” They only contain paracrine granules from what we have observed.

Nicolas Locker (00:55:03):
I guess it’s a virus that replicates very, very quickly. One round of infection is six to eight hours, and it induces… Which is the time at which we detect paracrine granule, and as soon as you go to 10 or 12 hours, you already have apoptosis and cell death, so we can observe from time to time the assembly of stress granules, but I think this is more condensation associated with basically the cells being on the way out and dying already rather than something biologically relevant.

Balaji Olety (00:55:45):
Thank you. I have a question. I guess this question was very obvious, right? So what’s the nature of this virus-free supernatant that you [crosstalk 00:55:56]?

Nicolas Locker (00:55:55):
Okay. Yeah. Sure. When we had this data, obviously we had a lot of people whether when I present, or reviewers of grants and papers being read out, that must be interferon. You have interferon in the supernatant, and that’s the mediator. Well, if we stick feline interferon on these feline cells and look at G3BP1 localization, nothing happens. Okay? This is not interferon-driven.

Nicolas Locker (00:56:31):
We thought about small RNAa, so we took the virus-free supernatant. I did recombinant RNase. The supernatant still had the ability to induce paracrine granule. Okay. What about DNA molecule? We used DNAse. Still had the ability to induce paracrine granule.

Nicolas Locker (00:56:52):
We know that paracrine signaling in the recent years has been proposed to involve small vesicles and exosomes, so we used antibodies against exosome to kind of potentially neutralize exosome vesicle in virus-free supernatant. We still had paracrine granule assembly.

Nicolas Locker (00:57:14):
The final thing that we did was to, based on studies, I believe, that come from… Well, I guess they must be coming from Roy’s lab showing that specific prostaglandin that are produced following translational inhibition are able to induce stress granule in a paracrine manner. We used ELISAs to quantify this specific prostaglandin and did not detect it in the virus-free supernatant, so not an exosome, not these other paracrine inducer, not RNA, not DNA. What does it leave us? Protein… Metabolites.

Nicolas Locker (00:58:01):
For the protein hypothesis, the only thing we know is that if we heat-treat the supernatant 60 degree for five minutes we lose the ability to induce paracrine granule, so it’s entirely possible that this could be a protein or small peptide.

Nicolas Locker (00:58:22):
Why don’t you use MS to identify that? Yes. So we’ve done mass spectrometry. We have a bunch of proteins coming up as present in the supernatant. It’s all the things that are released from infected cells, all the highly abundant junk that are present in the cell media that the mass spec people were pulling their hair out and complaining about us destroying their mass spec, so now we have at list of candidates that we have to go through, basically. I would hope that in maybe a year’s time or a little bit more we would be able to tell you a little bit more about the nature of this inducer.

Balaji Olety (00:59:10):
Thank you. That was classical virology, wasn’t it?

Nicolas Locker (00:59:13):
Yeah.

Balaji Olety (00:59:14):
Yeah. I mean, have you probably fractionated it?

Nicolas Locker (00:59:16):
Yeah.

Balaji Olety (00:59:19):
Have you excluded the possibility that it could be still virus-free but it could still be viral proteins that bind to the cell surface and induce some sort of signaling? Have you ruled that out?

Nicolas Locker (00:59:30):
Yeah. So what we did was we did some gel filtration to fractionate the extract, and we have narrowed it down to something that is between… That is around 10-30 kDa, but we haven’t been able to narrow it down more. In terms of viral proteins, from the mass spec we didn’t detect any viral product, and before doing that, we had… This virus only makes eight proteins. We have antibodies against all of them. We could not detect any of this viral protein in the supernatant, so we really do not think that this is a viral mediator but rather a cellular product secreted during infection.

Balaji Olety (01:00:31):
I see. Erik Martin has a question.

Erik Martin (01:00:34):
I just wanted to follow up quickly on that and ask if you’d looked for reactive oxygen species, or some kind of pH shift, or something as a result? Something very simple like that.

Nicolas Locker (01:00:49):
Yeah. Yeah. So we’ve been looking at reactive oxygen species, especially in the light of our recent data showing that stress granules can contain ROS. We did not. Simply by staining the supernatant with CellROX Green, and did not detect any ROS. When we look at the RNAseq we didn’t really notice any signature for ROS or mitochondrial stress and ROS-associated transcriptomic signatures.

Balaji Olety (01:01:38):
Thank you. To build on that, have you considered the possibility that this viral-free supernatant has a broad antiviral effect against other viruses?

Nicolas Locker (01:01:51):
Yeah. So that’s a very good point, and as a virologist that’s the first thing we did was to basically use the supernatant, stick it onto… First of all, we took the supernatant and obviously all the microscopy and all the -omics have been done in U2OS, so in human cells, but the supernatant is active in murine, human, feline. What else did we test? Porcine cell line. It has broad activities. Now, in terms of cross-virus reactivity/protection, we tested the other caliciviruses that we have in the lab and did notice a protection as well. We haven’t tested different families of viruses yet, so we haven’t tested DNA viruses or negative-sense RNA viruses like flu, or segmented viruses, so this is something that obviously we have on the agenda.

Balaji Olety (01:02:57):
Cool. Anymore questions? Well, thank you, Dr. Nicolas, for basically adding to the list of biomolecular condensates. This is exciting and fascinating as well. Thank you.

Jill Bouchard (01:03:14):
Yeah. Thank you also for all the questions for everybody that had questions for showing up. This has been super exciting. Thank you so much, Nic, for sharing all of your fascinating work with us.

Nicolas Locker (01:03:27):
My pleasure.

Jill Bouchard (01:03:27):
You can tell we’re all still very stimulated. Yeah. For anyone that’s still on the line, make sure to join us in about a month when we have our next Kitchen Table Talk, and thanks again, Nick, for joining us today.

Nicolas Locker (01:03:39):
My pleasure, and if anybody wants to follow up with questions, or if they drink their coffee tomorrow morning, and think about it, and have a super-cool idea for me, shoot me an email or message me on Twitter, and I’ll be happy to follow up.

Jill Bouchard (01:03:56):
Awesome.

Nicolas Locker (01:03:58):
My pleasure to talk to you, and if I’m ever in Boston in the next few months or years.

Jill Bouchard (01:04:06):
Come find us!

Nicolas Locker (01:04:06):
I’ll make sure I include a little trip to your kitchen table.

Jill Bouchard (01:04:12):
Sounds great. We’ll make sure to include your email address, then, when we post the video on condensates.com so anyone else can reach out. Cool. Well, thanks again, and thanks everyone for joining, and you all have a good day.

Nicolas Locker (01:04:23):
Cheers. Have a good day. Bye-bye.