VIDEO: Karla Neugebauer on Nascent RNA and Biomolecular Condensation in the Cell Nucleus

Karla Neugebauer (00:02:23):
I actually did this experiment at MPI-CBG, so underscoring my pleasure at being back here in Dresden. I picked it not because of that, but because it shows these thousands of transcription sites distributed throughout nucleoplasm. You can see even the nucleolus here is a big dark spot with a bright spot in the middle. That’s likely there for rDNA that’s being transcribed for nucleolus formation. I’m going to mention some work at transcription sites, especially in the beginning, because it informs how we’re thinking about Cajal bodies, which are shown on the right here.

Karla Neugebauer (00:03:02):
This is a recent STED image that instructed us anew as to the substructure now of Cajal bodies, which we didn’t know about previously. What we learned recently, and I’m going to be coming back to this, especially towards the end, is that the Cajal body, which we consider to be the green part of this lump, the Cajal body is marked by a scaffolding protein coilin, which I’m going to talk quite a lot about, is kind of like a baseball mitt or a ball in socket with another condensate called the gem.

Karla Neugebauer (00:03:37):
It now seems that these two condensates. I’m going to call them condensates, even though I actually don’t formally know they’re condensates, so my skepticism can come in towards the end. But anyway, that these two objects in the nucleus are very intimately related, but they’re not one and the same. So for example, the molecules that are present in gems are not detected in Cajal bodies by proteomics methods. I’m going to show you those data because we’re doing a lot of proteomics right now. That was a little bit of a preview. Let me move forward with some of the things I wanted to say.

Karla Neugebauer (00:04:14):
So first, I just wanted to point to the importance of nascent RNA in the nucleus. This is not a nucleus. What this is, it’s a very large pair of chromatids that have been dissected out of a newt oocyte. These oocytes are very highly transcriptionally active. In oocytes of amphibians and insects, different species, you can actually haul the chromosomes right out of the nucleus, and spread them out in buffer, and have a look at them in the light microscope. That’s what you’re looking at here. What you can see are these two chromosomes. This happens to be chromosome seven, and you see these lampbrush loops that we’re used to looking at occasionally in textbooks, but they remind us that the loops that form off of the chromatid axis are transcriptionally active genes basically.

Karla Neugebauer (00:05:10):
A lot of these loops are not recognizable as loops, but rather as balls of stuff. I would suggest that these balls are all condensates that represent the proteins and the RNA that accumulate at highly active transcription sites. It reminds us that in the nucleus, objects can form at highly active sites of transcription by virtue of the association of stuff in general with the nascent RNA that’s tethered to the chromosome access. You can imagine it like Velcro, because you’ve got the DNA as a substrate, and you’ve got these hair-like projections, which are the nascent RNA acting as a multivalent lightning rod or something like this. This is a backdrop as how I look at the dark matter of the nucleus. I don’t know why my slide’s not advancing. Let us… Here we go.

Karla Neugebauer (00:06:10):
This is what happens when you centrifuge those lampbrush chromosomes down on a slide. Now, you can fix and stain, and that’s what you’re seeing here. This is, again, to continue on this vein. Here are these transcriptionally active loops, some of which are very easily recognizable by eye, and so these early structures are actually genes that the cytologists recognized in late 1800s. Today, we might call those condensates.

Karla Neugebauer (00:06:37):
Up here on the left corner, you can see what were then called spheres. We would call these Cajal bodies the objects that we work on in the nucleus. So what was observed by Joe Gall and Mick Callan was that these Cajal bodies actually are nucleated on chromosomes on specific chromosomal sites. You can recognize them because they’re usually two balls at the same position on the sister chromatids. They can be nucleated on DNA. At the same time, you can form extra chromosomal versions of these in oocytes, because you’re providing the embryo with lots of stuff, like 1,000 nucleoli and 100 Cajal bodies. Based on our work, which I’ll mention, this is to support mRNA biogenesis in the early embryo. Here, what you can see on the right are some extra chromosomal Cajal bodies that have been seeded over here, and they’ve floated away, and they sort of dot the slide, therefore. Those are some opening comments. Let’s see. I’m not sure why I have to hit this so many times.

Karla Neugebauer (00:07:43):
So coming to the cartoon version: here, we have the nucleolus showing that rDNA is transcribed to form nucleoli. Here, we have a protein coding gene that we could view as a object or a location inside of cells. It might be different from the next gene. Then we have the Cajal body, which is the subject that we’re going to talk about today. There’s an interplay between the Cajal bodies and these sites of transcription and RNA processing of protein coding genes, because Cajal bodies support the assembly of spliceosomal snRNPs, which are required for the splicing of pre-mRNA transcript elsewhere in the nucleus, not in Cajal bodies, but at their transcription sites.

Karla Neugebauer (00:08:30):
I’m going to give a little bit of introductory information about the spliceosome, because that’ll be necessary for some of the things I’m going to mention. This is called the pre-mRNA. This is called the spliceosome cycle. I can’t see the title of the slide, because of the people at the top, but oh well. What you’re seeing here is a typical pre-mRNA diagram, which has two exons and an intron. What’s being shown is that sequentially recruit these different balls marked U1, U2, U4, U5, and U6. These are these spliceosomal snRNPs. It seems very complicated, and it is complicated. There’s 200 proteins. There are all these RNPs that I just mentioned, U1, U2, et cetera. They’re all going to associate with this pre-mRNA sequentially into at least 10–I think we’re up to 10 different transition states in the assembly of the spliceosome; we’ve got eight different helicases, splitting ATP every which way; we don’t actually know how many ATPs a spliceosome needs because they probably cleave ATP multiple times. Then we finally arrive to catalysis over here.

Karla Neugebauer (00:09:44):
The important thing to know for this talk is not about splicing catalysis, but rather the effect that splicing has on the snRNPs. In particular, I want to point to the U4, U5, and U6 snRNP that arrives here as a complex of three different small non-coding RNAs with a whole bunch of proteins. It’s been assembled in a very specific way. What happens at the end of splicing is it’s released, disassembled, and nonfunctional. So if we were doing a lot of splicing as neurons are doing or early embryos are doing, we could use up the spliceosomal subunits that build spliceosomes in a hurry by doing a lot of intron removal.

Karla Neugebauer (00:10:27):
So why would be doing a lot of intron removal? That’s because each gene in humans has at least 10 introns, and so what you could imagine here… Imagine this is a gene. This is a really bad diagram of a gene, because introns are 10 times longer than exons. These should be that much longer, but let’s forget about that detail. In fact, each one of these introns has to have a spliceosome built upon it. Just the same way I told you for that one intron example, in vivo, we really need 10 of these spliceosomes to assemble on every nascent RNA, whereas we only needed one measly Pol II to holoenzyme to transcribe that gene. It’s incredibly costly.

Karla Neugebauer (00:11:11):
This ribosome-sized machine has to assemble on each individual intron in order to get it out. You could think it could add time to gene expression to have to do this. But fortunately, most splicing is very efficient, and it’s even co-transcriptional. What that means is that splicing can generally complete in the same timeframe as transcription. That’s a relief. However, don’t forget that you need all of these splicing components to assemble these spliceosomes.

Karla Neugebauer (00:11:51):
This occurs in the Cajal body. So remember, I told you how these splicing components are taken apart at the end of splicing. What happens is they get recycled, of course, because the cell wouldn’t be so stupid as to throw away these little subunits that they could perfectly well reassemble. What happens is these subunits get transported to the Cajal body, which I’m about to tell you a lot about. That’s where this reassembly process takes place, so there’s de novo snRNP assembly there, but there’s also recycling that goes on there. It turns out this allows cells and indeed entire organisms to live. I’m about to show you the evidence for that in the next slide. I do not know why my slides do not advance, but I will keep pressing things. Aha, there we go.

Karla Neugebauer (00:12:39):
So while I was at MPI-CBG, Magdalena Strzelecka, a graduate student in the lab, did this fantastic experiment that enables me to tell you more also about Cajal bodies. So first of all, I can show you here is a green ball that’s supposed to represent a Cajal body. I actually haven’t shown you Cajal bodies before, so here now, I’m showing them to you. These are Cajal bodies present in the blastula cells of a zebrafish embryo. You can see this protein coilin–that I’m going to mention a lot–concentrated–this is one nucleus–concentrated in these round blobs. Here are the snRNPs, these guys that I’ve been talking about, concentrated also in these Cajal bodies together with the coilin.

Karla Neugebauer (00:13:22):
If we get rid of coilin using a morpholino, what happens is you lose the coilin, of course, but you also lose the concentration of the snRNPs in the Cajal bodies. You don’t lose the snRNPs, but you use their concentration in the Cajal bodies, where I just told you that we’d previously shown this assembly could take place using various methods like FRET methods and so on looking at transient intermediates in the snRNP assembly pathway. We can take immature snRNPs, and make them into mature snRNPs. This is how they look. What’s the result of knocking them out? The result of knocking them out is death.

Karla Neugebauer (00:14:03):
So the embryos die before 24 hours of development, and the reason they die is that they’re splicing deficits and a decrease in the number of assembled snRNPs. We can rescue them if we give them back snRNPs. The snRNPs don’t contain coilin, but they do contain this essential material that’s important for splicing. That’s the basis of our argument that the reason you need snRNPs to undergo embryogenesis is to support the removal of introns in the early transcripts made by the embryo. We believe that Cajal bodies are important, and we are really dedicated to trying to understand how they form and function. All right, great.

Karla Neugebauer (00:14:48):
What are the burning questions about Cajal bodies? How do they form? I just said that. Another question is do they have any other functions? That’s one of those questions you make up after you know the answer. It seems to us that they do have other functions, and I’ll allude to those in a second. Also, to come back to the question of what are the components of Cajal bodies, it turns out that most condensates or compartments in the nucleus are haphazardly described, and so I’m going to take you through an attempt to comprehensively define the components of our nuclear body.

Karla Neugebauer (00:15:25):
Yes? Is there a question?

Antonio Domingues (00:15:25):
I think somebody is- Yes.

Bede Portz (00:15:26):
Yeah, I have a question.

Karla Neugebauer (00:15:26):
Great.

Bede Portz (00:15:30):
So you say other functions. What is the primary function? Do you think Cajal bodies and coilin serve to chaperone the assembly or reassembly of snRNPs? Is that your thesis?

Karla Neugebauer (00:15:40):
Yes. So we… Sorry, I had to skip over that older work, but we did all kinds of Monte Carlo simulations of snRNP assembly. I’m looking at you in the screen. Sorry, that’s weird, isn’t it? We have shown that snRNPs assemble. This tri-snRNP that’s made of three different RNPs assembles in the Cajal body. We can observe this with FRET. We can observe the transient intermediates, and the outcome when you get rid of the Cajal body is that they don’t get concentrated into the Cajal body, and likely assemble at a much lower rate. That’s supported by the experiment that I just described to you where the embryos couldn’t live, because they couldn’t splice out their introns.

Bede Portz (00:16:27):
That’s a perfect segue to my follow-up question. You mentioned specifically neurons and embryogenesis. This is a naive question, but are there acutely high splicing demands in those tissues or developmental stages? Likewise, messing up Cajal bodies is embryonically lethal, but can you… What about like HeLa cells, for example, can they tolerate perturbation to Cajal bodies?

Karla Neugebauer (00:17:00):
That’s three questions you’re asking me. All right, great. In the embryo, you’re absolutely right. You need to splice the transcripts, and you desperately need spliceosomes, because you transcribe and process the RNA within a single cell cycle, which in embryos is 15 minutes long, at least in these embryos. They have very short cell cycles, and so our argument is that this potentiates the need for efficient assembly of snRNPs. In a HeLa cell, related to that, HeLa cells divide every 24 hours, plus they’re aneuploid, and they don’t even care if you hit them with all kinds of poisons.

Karla Neugebauer (00:17:38):
So, it is true. You can delete coilin from HeLa cells, and they seem to be just fine. I don’t really care that much about that. There are probably a lot of things HeLa cells can do without in order to divide, and be these horrible cancer cells. It does mean that you need to go to maybe more physiological or challenging situations that depend on efficiency. I’m not telling you the only way to make a snRNP is to go to a Cajal body. I’m telling you that at least according to our modeling and the tests that we’ve done, we think that assembly is more efficient in the Cajal body as a result of at like 20 fold–this is our estimate, a 20 fold increase in the concentration of the precursors that’ll now meet each other more efficiently in this location if you…

Karla Neugebauer (00:18:33):
The rate limiting step is actually the productive hit between two substrates. They have to hit each other by brownian motion. They have to find the face. It’s like a thousand fold hit that you take to the efficiency of assembly of anything, because you have to keep hitting the face of your partner, and find the guy that you’re going to actually bind to. That’s why we think concentration in a particular location is such an important thing.

Karla Neugebauer (00:19:03):
However, I do have a favorite idea, which is that the molecules within the Cajal body could establish this reactive surface for snRNPs. That would be additional to this idea of concentration. We would really like to ultimately visualize by molecular structure what’s going on inside. You’ll see that we’re getting there with single amino acid mutations and so on to see exactly what this object means for how the snRNPs are oriented. But I mean, this is actually a big macromolecular question. I’m not exactly sure how this is going to pan out. It’s a few steps away.

Karla Neugebauer (00:19:45):
Now, you also asked me about the brain, and I mentioned it because Cajal bodies were discovered by Ramón y Cajal in 1895 in the vertebrate brain. So one of the things that naturally happens to me in grant panels is people say, “Well, most cells don’t have Cajal bodies, and so they’re very weird, and therefore, this question is unimportant.” I want to tell you that every cell we’ve ever looked at in the zebrafish has them. Plant root cells have them. Brains of owls have them as Ramón y Cajal showed. The weirdos are aneuploid tissue culture cells that don’t divide very quickly. So cell lines that divide slowly do not show… We do not see them as large objects, if you get my drift. I think you know what I mean by this. We can come back and discuss these things more.

Jill Bouchard (00:20:37):
Which we highly encourage. Thanks, Karla.

Karla Neugebauer (00:20:41):
Now, burning… How do they form, and do they have other functions? I do not know why I cannot advance my slides. Does anyone have any suggestions?

Antonio Domingues (00:20:50):
No, it’s probably just slow Zoom.

Karla Neugebauer (00:20:56):
Aha! Okay. Here is a study that we published in 2014 that informed us on another function that we don’t really know…. we don’t know if it’s important, and so I’m welcoming comments. This was a study done by Martin Machyna, actually largely in Dresden, and then finished off at Yale. What he did, he kind of threw some omics at this protein coilin. This is, again, this scaffolding protein of Cajal bodies. If you get rid of coilin as we did in the zebrafish, he get rid of the object. What he did was he did a CLIP experiment. This is a UV crosslinking experiment where he asked, “Are there any RNAs that bind to this protein? And if so, where do they bind?”

Karla Neugebauer (00:21:39):
The answer was, yes, there are many RNAs. Of course, he wanted a long non-coding RNA, but he didn’t get any single long non-coding RNA. He didn’t get any mRNA. He got these small RNAs, U2 snRNA, U3 snRNA, U5 snRNAs. These are the RNAs that form those snRNPs. The coilin is actually seemingly binding because UV crossing is a zero angstrom method, they’re binding to these snRNAs, the coilin molecules. In the same breath, he did a ChIP experiment. So he said, “I wonder if I could ChIP coilin to any particular chromosomal sites,” and indeed he could ChIP them not to any other location, but to these genes encoding those snRNAs.

Karla Neugebauer (00:22:28):
These happened to be Pol II genes. These are Pol II driven individual genes that produce these small RNAs. They’re very tiny genes. They have transcription at them, and then here is the coilin pileup on top of these genes. That says that coilin is on the chromosome, and then it can bind RNA. The result of that was the drawing of this model that the coilin in the green ball here might be binding to the nascent RNA, and that’s why they form as a larger object on these sites, as visible by the coilin, and why they’re transcription dependent. I didn’t mention that previously, but in mitosis, they dissolve, and in transcription inhibitors, for example, they largely go away.

Karla Neugebauer (00:23:26):
This is also born out by an in situ hybridization experiment done by a different lab, by the Dundr lab, and so I just want to focus your attention on this. This is a multi-color FISH experiment. Here’s coilin in blue. That’s the Cajal body. Then here are four different loci, these U1, U11, histone two, locus–I’m sort of skipping over the histone relevance here–so very often, Cajal bodies are merged with what are called histone locus bodies into one object. And they’re very abundant on the histone genes in many of the systems that we work on. This cell, you can see all these gene loci clustered around the Cajal body. It’s not on top of the Cajal body, but adjacent to. We still consider this to be a working model, right? So we haven’t proven that it’s the nascent RNA that nucleates the Cajal body, but it seems very clear that the Cajal body has the capacity to pull together different chromosomal regions that represent these different genes, and these different genes are on different chromosomes.

Karla Neugebauer (00:24:34):
I didn’t say which chromosome, because in different species, there are different chromosomes, but these are independent chromosomes. So in other words, another function of Cajal bodies would be to pull together these different chromosomal arms bearing the U2 tandem repeat, the histone repeats that contain all of the replication dependent histones. U5, U3, these are all coming together at the Cajal body. We don’t know, again, if this chromosome conformation that’s created by the Cajal body is essential, but it’s certainly having this effect, and so it’s worth keeping this in the back of our minds as we try to entertain potential new functions for the Cajal body, certainly a morphological function.

Karla Neugebauer (00:25:26):
Those are the burning questions. How do they form? They form at transcription sites. Do they have other functions? Well, maybe this clustering function. We also know that they have a function in snoRNA assembly–I skipped over that, and let’s keep it that way. What are the components of Cajal bodies? These snRNA, snoRNAs, coilin as I’ve told you about it. Now, I want to delve in more detail this “how do they form” question. Let’s expand that too. How does this larger object form? You know that histone locus bodies probably are these little tiny dots that are really the transcription sites of histone genes. What it seems to us is the Cajal body is a bigger object, the larger object, and so it seems to be a higher-order object than simply a transcription site. This is what I mean by how does the larger object form. We also wonder about novel functional protein groups that could be in there in terms of comprehensive identification of components as well as the function question.

Karla Neugebauer (00:26:40):
Let’s start with what we’ve been learning recently about assembly. All right, so here’s… Finally, I’m telling you about the molecular structure of coilin. It’s a protein with about 580 amino acids, and it has an N-terminal domain, an intrinsically disordered region in the middle, and a C-terminal domain at the end. The C terminal domain is actually a folded Tudor domain that lacks the capacity of a normal Tudor domain, which is to bind dimethylarginine. I’m going to be talking about dimethylarginine in the last section of my talk. The reason I’m going to focus in the next few slides on the N-terminal domain is that it, like the C-terminal domain, is among the highest conserved portions of the coilin, and because the deletion of the N-terminal domain prevents the de novo formation or the localization of the molecule to existing Cajal bodies.

Karla Neugebauer (00:27:43):
So in other words, if I transfect a delta-NTD-construct into cells that have Cajal bodies, that protein will not go there. If I have cells that don’t have coilin, and I transfect a full-length coilin molecule, I can make new Cajal bodies, but not if I lack the N-terminal domain. That’s why we thought N-terminal domain is super important for assembly. I’m going to tell you about these experiments working on the N-terminal domain. This was started by Martin Machyna, the same person who did the CLIP and the ChIP experiments. What he decided he would do is start transfecting cells with only the N-terminal domain, and see what happened. What could he build up with just the N-terminal domain? This is the domain that we attribute assembly to.

Karla Neugebauer (00:28:35):
What he did was he took these coilin knockout cells–they’e mouse cells–and he expressed the NTD initially without a nuclear localization signal. He forgot the nuclear localization signal. What he discovered was these large fibrils or filaments or oligomers, or I don’t know what you want to call them. I’m arguing with the journal editors right now what should we be calling these things. They kind of look like C elegans embryos. Crazily, one of my grant reviewers thought they were C elegans embryos. They’re about 10 microns long, these guys here. They’re really impressive. We don’t know that they’re exclusively made of coilin. We have not been able to find any other component present in them, and we have shown that they’re not amyloids. The jury’s out what exactly is in them. If you express the NTD, however, in the nucleus, this does not happen. Instead, you get these little round puncta as well as some hazy signal in the nucleolus itself. Let’s ignore that for now, if you don’t mind. We get these little round things in the nucleus, yet we get these big, long things in the cytoplasm. What does that mean?

Karla Neugebauer (00:29:52):
So since we thought this was cool that the NTD shows the ability to self-assemble in a different way in the cytoplasm from the nucleus, and in the nucleus, it makes round things that resemble the round things that are Cajal bodies, we decided to move forward and look for amino acid changes that could modify those behaviors. So here, you can see the sequence of the N-terminal domain. You also see a predicted structure down here at the bottom, which is a ubiquitin-like fold. I’m going to show you that in just a second. You can see that these conserved amino acids often correlate with the presence of predicted structure, so these sheets here.

Karla Neugebauer (00:30:36):
This is an alanine scanning type approach that we took. Then what we did was transfect different constructs with these mutations, individual mutations, to see if any of these affected outcome. These are some of our favorite mutants. Here, what we’re doing is transfecting the NTD into the nucleus. Here’s the wildtype. Again, you see these little round dots. Remember, we’re going to ignore the nucleolar staining. These are actually HeLa cells, so you can see the endogenous coilin is overlapping with the NTD that has gone and joined the coilin here in the HeLa cell. Here’s the merge. Those are the nucleoli.

Karla Neugebauer (00:31:17):
All right, so here’s one of our favorite mutants, R8A. It makes the mut-NTD completely incapable of assembling into a round dot. As you can see, it’s dusky, distributed all over the place. Not only that, it takes apart the existing Cajal bodies, so it’s got a dominant negative effect just like its friend D79A down here, which is also dispersed throughout the nucleus, and it has taken apart the Cajal bodies, so lots of nice examples of dominant negative mutations. Perhaps as we believe related to the fact that these NTD molecules can interact with each other, at least as dimers and potentially as some other form of multimers that’s behind the assembly of the Cajal body.

Karla Neugebauer (00:32:07):
This is actually my favorite mutation, R36A, which rather unbelievably causes the NTD to now form filaments very similar to those cytoplasmic filaments that we showed, and these filaments form in the nucleus. Ha! What I’m going to suggest is that what’s happened here is we have disrupted interactions between the coilin NTD and a ligand in the nucleus. The reason we made filaments out in the cytoplasm is that that ligand is not present in the cytoplasm. So in other words, the R36A mutation in the nucleus is very similar to putting this protein in the cytoplasm where the ligand does not exist. Not only that, but once again, it’s a dominant negative mutation. Look at that. It has sucked up the endogenous coilin wildtype molecules onto these oligomers or filaments in the nucleus, again, suggesting that these molecules can associate with one another into higher-order structures. You can take these wildtype dudes, and bring them into these aberrant structures. No problem. Yes?

Antonio Domingues (00:33:16):
Can I ask myself a question?

Karla Neugebauer (00:33:18):
Yes. Antonio has a question.

Antonio Domingues (00:33:21):
It seems like, especially in the mutation, the R36A, that there’s a very clear… They are together, but there’s a very clear distinction. Half of the fibril is from the coilin, and the other one is from the other one.

Karla Neugebauer (00:33:33):
That’s a great observation. There’s some kind of heterogeneity here in terms of, exactly, the preponderance of the red versus the green, so that’s the HA tag versus the endogenous coilin. Exactly, we don’t understand the structure of this thing yet, so I think you’re right that it’s a patterning. There’s some kind of patterning going on, and we don’t understand this quite yet at the molecular level. We really need to do molecular structure on this. So obviously, we can benefit from these mutations in doing structure. I’ve got a new student in the lab who’s hell-bent on biophysics here. So, when I advance the slide, you’ll get a flavor for how we’re viewing this.

Karla Neugebauer (00:34:23):
Here’s a predicted structure. Remember, I showed you those sheets, and there was a coil. This is the kind of ubiquitin-like fold that’s predicted. It turns out these amino acids, interestingly enough, are on different faces of one another, so you can imagine all kinds of models for how coilin-coilin interactions look versus coilin-Nopp140, -something else interactions. I already named it Nopp140. Exactly. So we… This is the name of the molecule that we guessed it might be, and we have really good evidence that it is Nopp140. I’ll take you through that. Look at the arrangement here. So these amino acids that we’ve implicated in coilin-coilin interactions based on these mutant phenotypes are on the one face, and amino acid that we’ve implicated in this filament or antagonizing the filament formation are on the other face. Why did I whip out on the name Nopp140? It’s not a very attractive name even. I wish it had another name.

Karla Neugebauer (00:35:32):
Actually, the gene name is NOLC1, which will become more important in a minute. The reason is that there’s an very old paper that used the yeast 2-hybrid type of strategy to say that coilin interacted with Nopp140. It was on our candidate list. Nopp140 is actually an abundant protein in nucleoli. There are some nucleolar components that are present in Cajal bodies, so we knew about this. Also, just staring a Nopp140 is just amazing, because it’s 700 amino acids basically of intrinsically disordered region. The only structured region is this pathetically small LisH domain at the end terminus, so it’s just not very interesting there. I’m going to show you in a minute that it’s capable of condensation, which is completely not surprising.

Karla Neugebauer (00:36:24):
That’s why we got interested in Nopp140, and we predicted then that coilin will bind directly to Nopp140. We could confirm that using a FRET assay. So on top here, this is an indirect FRET in cells that looks at coilin-Nopp140 on the one hand, so these are CFP-YFP pairs. And then we looked at coilin-coilin FRET down below. In the Nopp140 FRET, you can see that this R36A mutation completely abolishes the interaction between coilin and Nopp140, which we punitively assigned to this interaction face. Then coilin-coilin FRET is completely abolished when we mutate R8A over here on this other face. There are clearly more complicated things going on here, because some of these mutations don’t play out exactly as how we predicted it, but still, the results are pretty consistent with these ideas that we have coilin-coilin interactions on the one side, and coilin-Nopp140 interactions on the other side.

Karla Neugebauer (00:37:29):
Then finally, we used… I’m going to talk about this later, but the CRY2 assay from Cliff Brangwynne’s lab, I’m imagining that this audience is pretty familiar with the Cry2 assay. I threw this in here just to show you the result. First of all, on the right, this is the evidence that Nopp140 can do biomolecular condensation as defined by the Cry2 assay of Cliff Brangwynne again, where when you express this construct, we took the IDR of Nopp140, and hooked it up mCherry-Cry2, you express it in 3T3 cells. You see very uniform but high expression. Then when you turn the light on, and dimerize the Nopp140 construct, you see these very large, dark, condensate type objects. Then of course, when you turn the light off, they disperse, and so that’s our definition of condensation in this context.

Karla Neugebauer (00:38:26):
Nopp140 fits the bill. You could imagine Nopp140 then being the IDR providing condensation function for coilin in trans, because coilin is binding to this. Coilin itself, you could say, “Well, coilin has an IDR. Why does it need to go hang on to someone else?” But when we take the coilin IDR, and make a construct out of that… Basically, what I’m showing you here, I just chose to show you the delta-NTD construct. That’s the coilin IDR together with the C-terminus. You can see this even expression again, very high level of expression, turn on the light, and no condensates are forming. We basically don’t have any evidence that any portion of the coilin molecule has the ability to make condensates, at least not in this assay. That’s why we have proposed that coilin binds to Nopp140, and does so in trans.

Karla Neugebauer (00:39:22):
This is our little model then that posits that out in the cytoplasm. These amino acids that we’ve been able to study through these single amino acid mutations can mediate fibril formation, but in the nucleus, this is buffered by binding two Nopp140, these little aqua structures here that either limit oligomerization by coilin, or take long oligomers, and maybe even wrap them up into a ball like a ball of yarn. We don’t know which of those two hypotheses is correct currently.

Karla Neugebauer (00:40:01):
I just told you what we know about coilin’s role in the assembly of the Cajal body. Now, I’m about to launch into a discussion of our mass spectrometry study, where we go to de novo identify Cajal body proteins. I’m going to be telling you about some new constituents. Does anybody have any questions about that assembly part that I talked about?

Antonio Domingues (00:40:30):
I don’t know if it’s a question about the assembly part, but Jon Henninger has a question.

Jon Henninger (00:40:34):
Yes, actually it is a question about the assembly part. So as you’re probably aware, Tony Hyman and Simon Alberti had this paper in 2018 about the phase separation of FUS and the nucleus versus the cytoplasm, and how the total RNA levels actually regulate that. It struck me that most of your mutations are in arginines. I’m just wondering if you think that RNA could be mediating this behavior.

Karla Neugebauer (00:40:58):
Exactly. Thank you for asking that question. We have… Of course. There are different kinds of RNA in the different compartments. I’m sure everyone knows this. The snRNAs, for example, are at high concentration in the nucleus, and they’re essentially absent from the cytoplasm, same with the snoRNAs. These are the relevant RNAs for Cajal bodies, and so… But at the same time, there could be a buffering effect of RNA in general. I didn’t mention that aspect, because I just don’t know what to say about it at this point.

Karla Neugebauer (00:41:31):
I have shown you that coilin binds to RNA due to this CLIP experiment, but we don’t know what region in coilin binds to RNA. That’s part of the reason we picked these arginine molecules. Officially speaking, there could be, for example, an RNA between the two NTDs that are binding to each other, and that we’re detecting by FRET. I mean, FRET is a very close interaction, so it’s possible that there’s an RNA involved, but we currently don’t have any evidence for that. What we’re doing right now is we’re trying to really rigorously define where the RNA binding domain on the coilin protein is.

Karla Neugebauer (00:42:16):
I’m sorry that I can’t tell you more about that, but we do suspect that it’s actually further down towards the C-terminus where there’s an RG repeat, and there’s other indications that the snRNPs actually bind towards the C-terminus. So basically, all the suggestions point to this more C-terminal region is binding the RNA, but we haven’t rigorously shown that yet. I hope that answers your question. Now I’m going to tell you about using coilin to identify the proteins in Cajal bodies.

Jill Bouchard (00:42:51):
Karla?

Karla Neugebauer (00:42:52):
Yeah.

Jill Bouchard (00:42:52):
We have a Kitchen Table question.

Karla Neugebauer (00:42:56):
Another one. Awesome.

Tina Han (00:42:57):
I have a question about the coilin fibrils. Do you envision any function for the fibrils?

Karla Neugebauer (00:43:05):
Well, we haven’t seen fibrils. No one has seen any fibril aside from this unusual situation that the NTD is expressed in the cytoplasm. In fact, when we express the full-length protein or GFP-tagged NTD in the cytoplasm, we don’t get them. I don’t think that the fibril formation in the cytoplasm… It shows that the NTD can do this, but it doesn’t tell us when or if this is a physiological state. We’re seeing it as an indication that the NTD can do fibril formation, which we didn’t know before, but we’re not exactly sure if in the Cajal body, for example, if we do CryoET, would we see wrapped up fibrils?

Karla Neugebauer (00:43:54):
There’s actually… I didn’t put a slide in my thing. I could potentially maybe at the end go and get a picture if you’d like to see it, but old EM pictures of the Cajal body do… It used to be called the coiled body, because it looks like an electron dense coiled up thing. I’m very cautious about saying that these are completely artifactual, because it is possible that there are oligomers that are rolled up in the Cajal body, maybe like a jelly roll with the Nopp140 molecule, making it into a round thing, because it’s clearly… It’s never been seen before in the nucleus as a fibril.

Tina Han (00:44:37):
And then just a second question, did you say that… Do you ever see Cajal bodies in the cytoplasm?

Karla Neugebauer (00:44:47):
No.

Tina Han (00:44:48):
Never?

Karla Neugebauer (00:44:49):
No.

Tina Han (00:44:49):
Not even in pathological states?

Karla Neugebauer (00:44:52):
No. There is one pathological state related to the SMN protein that I’m going to talk about later. So this protein, when it’s absent, causes spinal muscular atrophy, which is degenerate motor neuron disease of children. It’s fatal. In patients with SMA, their Cajal bodies look abnormal, which makes sense; that SMN protein nucleates those gems that I showed you on the very first slide. I’m going to be coming back to that later. So again, that’s a case where even though SMN has a cytoplasmic role in snRNP assembly, no. No. No. Coilin’s never in the cytoplasm that I know of. I mean, obviously.

Tina Han (00:45:37):
Thanks.

Karla Neugebauer (00:45:41):
All right, so here’s… More questions?

Bede Portz (00:45:43):
I have a follow up to Tina’s question actually. So in the instances where you get these cytoplasmic fibrils… Apologies if I missed this. If you make point mutations that abrogate the ability to fibrilize, do you then… If those mutants are made in the context of the full-length protein, do they mess up nuclear function? In other words, is that fibrilization activity reporting on some functional self-assembly that even in the endogenous context is important?

Karla Neugebauer (00:46:18):
Well, kind of. If we look at these mutations, we would have to say that these amino acids are involved in forming fibrils, because when we get rid of those amino acids, these arginines, we now have just diffuse protein–doesn’t assemble into anything, and it also–which I think is very important–takes apart existing oligomers that were present, the wildtype coilin. I would think that these are the amino acids required for making fibrils. What the R36A is doing is allowing you to see those fibrils because it’s inhibiting binding to Nopp140. Do you see what I mean? I don’t know exactly… I mean, I guess what you’re asking for is the R36A mutation in the same NTD as we’ve made in R8A mutation. That’s a cute idea. We haven’t tried that. Yes, you’re right. We should do that. You’re right. We should do that. Awesome. Come join my lab. As you can imagine, there’s a lot of these permutations not to mention.

Karla Neugebauer (00:47:41):
Let me tell you about this mass spec experiment. Here’s what we did. You guys probably know Anne Gingras’ experiments, where she is using bioID to try to do comprehensive proteomics on different biomolecular condensates in cells. We took a slightly different approach to hers. We’re using APEX2, which enables us to do a one-minute peroxide treatment with the biotin phenol in order to biotinylate proteins that are in the neighborhood of Cajal bodies.

Karla Neugebauer (00:48:16):
That’s in contrast to… We started planning this before Turbo BioID, but, okay. A lot of the proximity biotinylation experiments that have been published have been done with much longer incubations of the appropriate reagent. We wanted to make sure that the incubation be really short, because we know that coilin actually leaves the Cajal body with a time constant of something like a minute. What we want is to have Cajal body sitting there, add biotin phenol and peroxide for one minute, and basically get a picture of what are the proteins that are near coilin in the Cajal body.

Karla Neugebauer (00:48:58):
What Dahyana Arias Escayola did was she made these beautiful cell lines, where she titrated extremely carefully the coilin APEX2 expression so that it was not expressed throughout the nucleoplasm, which you’ll see often in some high throughput proximity biotinylation type experiments. I don’t want to throw anybody under the bus, but if you see the whole nucleus is labeled, then you’re probably not only labeling the Cajal body. These coilin APEX2 constructs are in Cajal bodies, overlapping with coilin, and then they have snRNPs in them, and Nopp140 in them as you might expect. When we do the biotinylation reaction, what you can see is the biotin is actually localized in the Cajal body. This is what we wanted.

Karla Neugebauer (00:49:54):
Now, we have controls, and our two controls are the delta NTD construct. So again, remember, I showed you the Cry2 experiment. This is exactly the same result. If we cut the NTD off of coilin, and have it connected to APEX2, it’s expressed all over the nucleus, and so is the biotin created there. Then the APEX2 NLS construct is simply, again, to go after any background type of signal that you could imagine getting, right? It’s very hard to make the right controls for these types of experiments, but we’re doing our best. I think these are two pretty good ways of looking at the data.

Karla Neugebauer (00:50:35):
We collaborated with Falk Butter and Emily Nischwitz in Mainz, old colleagues of Antonio, and also old colleagues of ours. We’ve done numerous mass spec experiments with them. Falk is just an amazing scientist. He does mass spec, but he’s also really great at the data analysis, so I feel very confident when we come up with these volcano plots, where you can see some old friends. These are known in orange. There’s coilin. There’s TGS1 always is the top hit in our hands. That’s an enzyme that puts the cap onto snRNAs. Then here is NOLC1 one. That’s Nopp140. That makes a lot of sense that Nopp140 would be labeled by coilin.

Karla Neugebauer (00:51:26):
There are also these green dots, so many new components. We had 100 proteins enriched. And of these, we would call 70 new components. I’m not going to show you all their names, and talk extensively about each of their roles, but what I would like to show you is a summary table that puts them into categories based on what they do. We have DNA replication repair, cell cycle progression and transcription as major new hits.

Karla Neugebauer (00:51:58):
So if you compare that to what was known before and what’s new, this is really upping the number of factors that are involved in transcription and DNA biology as compared to the other categories, for example, pre-mRNA splicing and 3 prime end cleavage. Many of these were already known. We’re adding a few new components, but this was already the most highly enriched category that we had before. What we’ve got here is a couple of new areas. In particular, I think it’s striking that now, transcription is that much more elevated that it’s on a par with splicing factors. This goes back to our notion that the Cajal bodies are nucleated co-transcriptionally, and so we need to dig into maybe some of the transcriptional relevance more in the future.

Karla Neugebauer (00:52:56):
The reason I’ve encircled these two categories in orange is these are involved in our rRNA processing, which we already expected, because I told you that snoRNPs transit through Cajal bodies on the way to the nucleolus (I think I mentioned that briefly). So we expected to see this, but what we really didn’t want to see is ribosomal proteins, because then you tell yourself, “Oh, this is just some kind of background or something like that.” But okay, we’re being comprehensive, so we’re going to keep them on the list. They passed all the tests, all the different controls and so on.

Karla Neugebauer (00:53:31):
Now what we did was we came to a functional test. So what we did was we did an siRNA screens in a high throughput format at the Yale Center for Molecular Discovery. Then we did a staining assay where we visualized nucleoli and Cajal bodies. We just asked for things that would disrupt the number of Cajal bodies. We’ve got all of the hits. We also have all the previously curated protein components of Cajal bodies that we could get out of the literature and so on.

Karla Neugebauer (00:54:03):
This is what we got. We got a number of… Let me see. What should I point at? I’m going to just point at this histogram here. What we got–which we expressed in this histogram as a normalized percent effect on the number of bodies, which were automatically counted by the imaging software–is we got 25 hits that decreased the number of Cajal bodies. Ok, Fine. You would think: you knock down coilin, you have less Cajal bodies. You knock down Nopp140, you have less Cajal bodies. This is all true. What was very striking is we got nine hits that increased the number of Cajal bodies, which is really weird, and so we decided to focus on these guys for the purpose of this study.

Karla Neugebauer (00:54:50):
Let’s pursue who those are. It turned out they were these components of the large subunit of the ribosome. Here, you can see these intrinsically disordered proteins that actually fold when they lie down on the ribosomal RNA like they’re on a beach or something like that. They associate with the RNA. It’s actually very beautiful. Here they are. They’re kind of clustered in a particular part of the large subunit. These are the two faces of the ribosome. These are the dudes who caused there to be more Cajal bodies, so just in a group. These are all involved, either involved in large ribosome subunit assembly or in the large subunit.

Karla Neugebauer (00:55:36):
What do I want to show you? One of the ways that you could imagine getting this is if the cells become aneuploid. We do know that Cajal bodies increase when you increase the ploidy, and that, again, has to do with the chromosomal association that you would expect. One of the hits is ANLN that actually blocks cell division, and makes cells become aneuploid. You can see that in quantification of the DNA content, ANLN knockdown is making cells aneuploid, and it also makes the number of Cajal bodies increase. But in contrast, testing all of these large ribosomal subunit proteins like RPL14, which I’m showing here, does not change the ploidy of the cells, so it’s not about ploidy. There’s something that happens when you knock down these particular only large subunit ribosomal proteins that causes there to be about twice as many Cajal bodies as there were before.

Karla Neugebauer (00:56:38):
We decided to apply this ChIP experiment that I told you about where we can ChIP coilin onto particular genes. I’m just showing you the histone gene cluster on chromosome six here. Here are all these little histone genes. You can see that in the control situation, the control siRNA, we see accumulation of coilin over these histone loci. Interestingly in the RPL14 and RPL24 knockdown, these become less, so the concentration over these particular genes is disrupted. It’s just a lot more like chatter.

Karla Neugebauer (00:57:20):
A zoom in to this particular gene underscores this, where you see a peak over here that resembles Pol II, for example, so this concentration of coilin that coincides with Pol II, in the two knockdowns is destroyed. Interestingly, the transcription of the histone gene is also… It’s suggested that the transcription of this gene is disrupted. This result actually implies that the extra Cajal bodies that we’re seeing are extra chromosomal, because they’re less well associated with the chromosomal regions than they were before.

Karla Neugebauer (00:58:10):
I want to give you just one more hint about potentially what these proteins could be doing. That comes from this imaging experiment that’s shown here on the left, where I’m showing you the coilin together with SMN staining. I mentioned this SMN protein before as marking the gems. So, what happens when you knock… This is showing this ball in socket arrangement or the baseball in the glove, where we see the gem associated with the Cajal body in the confocal microscope. Then when we knock down this large ribosomal protein, what happens is the SMN merges with the coilin. This is really a very distinct phenotype, and we’ve seen it one time before. I’m about to tell you about that situation.

Karla Neugebauer (00:59:06):
So first, I want to show you what this looks like in the STED microscope, because this is really a very dramatic phenotype. That’s why we believe that this is telling us something, even though we don’t wholly understand it. Here, what you can see is, again, the coilin marked in pink, and the SMN marked in blue. What happens when you knock down these two different ribosomal proteins is now, the SMN is mixing together with the coilin. We are changing the sub-structure of the Cajal body. The consequences of this for function, we’re not sure about, but obviously, we would be interested in learning what that is. We can show this with all sorts of quantification in terms of the offset and the intensity overlap, so we really believe in this phenotype as being important.

Karla Neugebauer (01:00:00):
Part of the reason we’re really gobsmacked to see this is that we had just seen this quite recently in a previous study, and that had to do with the study in which we described this arrangement. I want to now tell you a bit about SMN and its role in the assembly of the Cajal body, which, again, you see here by STED. This is the image of collected image of 20 different Cajal bodies reconstructed in three dimensions.

Karla Neugebauer (01:00:35):
You can see that it has this open structure where it holds the Cajal body… the gem, sorry, like in the baseball mitt. What I want to tell you about is how SMN actually has a function which is to bind this modification on arginine. This is an arginine side chain that you’re looking at here. This can be modified by dimethylation. The dimethylation can either be symmetrical as shown here. So in other words, each one of these terminal nitrogens gets a methyl group, or it could be asymmetrical in which case, one of the nitrogens gets two methyl groups.

Karla Neugebauer (01:01:21):
These are two post-translational modifications on arginine that have been studied in the context of stress granules. Some of you may be aware G3BP1 is modified by dimethylarginine. This is important for intramolecular interactions for G3BP1 in forming stress granules. What we found is that when we used aDMA inhibitors in our STED experiment, again, looking at SMN and coilin just as I showed you before for the RPL. When you use aDMA inhibitor–This is an inhibitor of the methyltransferase that makes this post-translational modification–you have a mixing of the SMN again with the coilin. This looks identical to this RPL protein knockdown. This is just astonishingly similar. The ribosomal proteins are not only intrinsically disordered, but they’re very heavily dimethylated, so we think that there’s some connection here, but we’re not exactly sure what it means.

Karla Neugebauer (01:02:30):
Let me tell you more about what we think is going on with SMN. This is actually what these… This arrangement, the ball in socket or the baseball mitt, looks like in the standard confocal microscopy. I showed you another image that looks like this, so I think you’re all used to it now. So here’s your coilin and SMN. I showed you this exact same… This is the same images I showed you before. This is with the aDMA inhibitor. Then over here, you see what happens with the sDMA inhibitor. It’s just something very specific going on with dimethylation, because here what you see happens is the gem completely separates from Cajal bodies. Both objects are okay. They’re just completely separated from one another. Have a look at this nucleus. There’s a bunch of pink dots and a bunch of green dots, and they’re not talking to each other anymore. So weird.

Karla Neugebauer (01:03:25):
Now, if we combine the two methyltransferase inhibitors, what you see is the gems are okay. If anything, they look a bit big. The Cajal bodies get really little, teeny, tiny dots. You could also say they disassemble. They just turn into these little wimpy tiny things. So, dimethylation’s really important for Cajal body integrity. We know that SMN binds to dimethylarginine.

Karla Neugebauer (01:03:52):
This is the molecular structure of SMN. It has a Tudor domain that binds dimethylarginine. So what is going on here? We think that SMN in the gem must bind at the interface, some molecules in the Cajal body part. We know that it binds coilin. I’ll show you that in a minute, but it binds coilin because we know the arginines and coilin are dimethylated. At least some of the arginines are dimethylated. How do we know that the Tudor domain is important for forming bodies or condensing at all?

Karla Neugebauer (01:04:31):
Well, we had this paper last year, and I’m… I don’t know. Maybe you read it. I’ll just remind you. This is, again, the CRY2 assay of Cliff Brangwynne’s lab. We used as a test domain the SMN Tudor domain. What we found in this study is that when you turn the lights on, now it forms condensates both in the nucleus and in the cytoplasm. The reason it’s present in the cytoplasm is it does some of the snRNP’s cytoplasmic part of snRNP assembly out there in the cytoplasm. It’s interacting with the snRNPs. The question is, what is it doing in the nucleus? Mind you, nobody knows what the function of SMN in the nucleus is.

Karla Neugebauer (01:05:11):
This is just a tiny bit that I took out of the paper. This is showing the molecular structure of the Tudor domain. There’s an aromatic cage, these green amino acids here that binds to the sDMA, which is fitting into this aromatic pocket. Here is one of the SMA mutations from a patient. This glutamate is hydrogen bonding to the back face of this aromatic cage, and when you mutate it to a lysine, it will not stab… It will destabilize the aromatic cage, so it binds less. It’s been shown in this paper. It binds less well to dimethylarginine, and causes a fatal childhood disease. So it’s important that SMN binds to dimethylarginine.

Karla Neugebauer (01:05:59):
We could show that in the absence of binding to dimethylarginine–either by mutating any one of these amino acids in the aromatic cage, or by using the same drugs I showed you befor–that we could block optoDroplet formation as measured by a clustering metric that we came up with alongside Joerg Bewersdorf, our collaborator. I feel like I’ve just said a whole lot of really long sentences, so I hope that was digestible to the audience. Maybe if you need any clarifying questions, you can go ahead and ask them. Are there any questions before I start to tell you about more ligands for the SMN-Tudor domain?

Karla Neugebauer (01:06:44):
Just a review, we’ve shown the SMN-Tudor domain is important for biomolecular condensation. We know the SMN-Tudor domain binds dimethylarginine modified proteins, but my argument is we don’t actually know what all of the ligands are for the SMN-Tudor domain. The reason that I don’t think we know what they all are is that coilins, dimethylarginine does not account for the creation of these optoDroplets. For example, we can knock down coilin, and SMN will still make optodroplets. That’s part of our set of arguments. Does anybody have any questions?

Antonio Domingues (01:07:25):
Not in chat.

Karla Neugebauer (01:07:26):
Are they… Do you think they are comatose at this point?

Antonio Domingues (01:07:29):
No, I think they’re very excited. I think there’s a number of questions coming up at the end, but we’re just holding off.

Karla Neugebauer (01:07:36):
This is my summary slide, so just highlighting the binding to transcription sites, the fact that it’s in a complex with gems, but that unlike the previous idea that the gem components were mixed with the Cajal body, we now realize that they’re not mixed, that they’re docked onto one another, and that that’s important for generating the Cajal body structure, that there are ligands for SMN that likely participate in generating Cajal bodies, and that we’re trying to find those and other RNAs that assist in this process. I also told you about a comprehensive discovery of these new proteins by this proximity method.

Karla Neugebauer (01:08:31):
I’m just going to quickly thank my lab. Then I’ll come back to the summary slide. This is Daniela, who’s doing the SMN-Tudor experiment. This is Leo with the idea to digest the extract with RNAs. This is Dahyana who did the APEX-coilin experiment and the functional screen. Unfortunately, Martin is… Oh, there is Sara Gelles. She’s helping out with the NTD project doing amazing work to try to understand these amino acid mutations from a biophysical angle. So, I’ll just go back to the summary slide, and see if there are any questions.

Antonio Domingues (01:09:12):
Right so-

Jill Bouchard (01:09:13):
Great talk. Thank you.

Karla Neugebauer (01:09:19):
What time is it? Is it midnight yet?

Antonio Domingues (01:09:22):
We’re over time, but it’s okay.

Jill Bouchard (01:09:23):
We’re a little over, but we definitely held you back a little.

Antonio Domingues (01:09:28):
I have a couple of questions, but I know that Bede, which is there also has a few questions.

Jill Bouchard (01:09:34):
I’m sure he has one too.

Antonio Domingues (01:09:39):
Do you want to start?

Karla Neugebauer (01:09:39):
Should I stop sharing? Do we want to talk, or does anybody want to go back to slides to see something that they-

Jill Bouchard (01:09:45):
Maybe let’s check with Shruti first.

Antonio Domingues (01:09:47):
Sure. Oh yeah. Sorry. I think it’s better to keep the slides for now in case there’s a question about the data. Shruti, do you want to go ahead and ask a question?

Shruti Jha (01:09:56):
Yeah. Amazing talk, probably one of the best that I’ve heard in a while.

Karla Neugebauer (01:10:02):
Thank you. Oh my God. That’s very kind.

Shruti Jha (01:10:03):
You’re so funny.

Karla Neugebauer (01:10:04):
Oh, I can’t hear you now. What happened? We don’t hear you. Oh, there you go.

Shruti Jha (01:10:18):
Oh, great. You had so many unsettling data, a lot. One of the most was-

Karla Neugebauer (01:10:27):
I’m sorry.

Shruti Jha (01:10:27):
The most unsettling was how you knocked down the ribosomal protein, and then you see the merge of Cajal and gem. What do you think is happening? Is it just methylation associated, how they are linked? Do you have any idea?

Karla Neugebauer (01:10:44):
One of the woo-woo ideas that we suggested in the manuscript is that these RPLs, we think about them as being on the ribosome. We’ve seen them in the molecular structure of the ribosome, but that they’re not always on the ribosome, right? So in the nucleus, you have the nucleolus, where the pre-ribosomal subunits are formed, so it… We do know that there’s a lot of exchange between the nucleolus and the nucleoplasm and the Cajal body. What we are wondering is whether… There’s a lot of possibilities, but it could be… There could be a condensate maybe made out of these intrinsically disordered proteins from the ribosome.

Karla Neugebauer (01:11:35):
I don’t know if I want to call it a condensate or a complex that might be present in the nucleoplasm, and interact with, let’s say, SMN in either the formation of the gem or the Cajal body, or buffer those interactions. Then if it’s not there, then you have forced SMN to start binding other ligands that might be present in the Cajal body. Like, why do these molecular entities invade each other now? That’s the question. So in other words, because each one depleted alone does it. That’s why invoked a complex, because otherwise, how could it be that they individually buffer something with their dimethylarginines?

Karla Neugebauer (01:12:26):
Do you see what I’m saying? If that were the case, wouldn’t we need to knock down all of them in order to see such a dramatic effect? That’s why I invoked a complex, and I’m talking to ribosome assembly type people to ask them if they know about this. I talked to [Rica Chotai], who’s a big expert in this.

Shruti Jha (01:12:48):
Yeah, she’s awesome.

Karla Neugebauer (01:12:54):
She thinks it’s an exciting idea, but she doesn’t know so.

Shruti Jha (01:12:56):
You think there’s a differential ligand in gem or in Cajal that’s causing that separation, but when you knock down RPL, or do some mutation, this difference does not exist, and they merge.

Karla Neugebauer (01:13:14):
Exactly.

Shruti Jha (01:13:15):
Cool.

Karla Neugebauer (01:13:15):
Right. I mean, the other thing to notice though is… I argued that they’re off the chromatin. Now, you could argue the reason they’re off the chromatin. It’s a cause or effect thing. It could be off the chromatin because for whatever reason, the depletion of these proteins is shutting down transcription, but yet now you still have the Cajal body formed. It’s just floating around, but my point is that’s something that’s also different. So if you look at my little model slide here, for example, I’ve got the coilin containing part on the nascent transcripts.

Karla Neugebauer (01:13:51):
But what I’m saying is now when we knock down RPL, maybe this is severed, and maybe it’s the RNA that’s buffering that interaction between the gem and the coilin. In the absence of transcription now, we can fuse the two bodies together. There’s a lot of different ways that you can turn around the evidence that we have. We have a lot of different symptoms of what’s going on, but we don’t really know the cause, which is really frustrating actually.

Antonio Domingues (01:14:30):
We’re way over time, so I’ll allow one more-

Karla Neugebauer (01:14:31):
Are we? I have no idea what time is it?

Antonio Domingues (01:14:35):
… one more question, quick question if anyone has one. There’s one.

Jill Bouchard (01:14:41):
I know Bede is brimming.

Karla Neugebauer (01:14:41):
There’s the APEX–Yes.

Jill Bouchard (01:14:42):
We actually have one from the audience.

Antonio Domingues (01:14:43):
There’s one in chat. Let’s-

Karla Neugebauer (01:14:44):
Leonard, yes, I see that. Exactly. That’s what we want to do now is the SMN APEX2, because… Thank you for bringing that up. This is how distinct these bodies are normally. We did the coilin APEX2. We got a lot of expected components, but I’m going to tell you, we didn’t get SMN. We didn’t get anything in the gem, which is the gemens. These things are distinct. We did… SMN does not ChIP to the loci. So even though these things are intimately associated, they act like different objects. Isn’t that amazing? We did not get SMN by apex coilin.

Antonio Domingues (01:15:28):
What I was wondering, I think that Leonard’s direction is when you get those fused, does that change? Do you start getting gemens, and you start getting–?

Karla Neugebauer (01:15:35):
Oh, is that… I can see the whole question actually.

Antonio Domingues (01:15:36):
The question is: have you thought about APEX proximity labeling with SM1 or SM1-Tudor domain fusions in comparison to the-

Karla Neugebauer (01:15:44):
Yes. Yes. Yes. First of all, we haven’t done SMN APEX2 at all. By the way, we did everything in the nucleus, so we’ve got all the conditions worked out so that we won’t get the cytoplasmic partners. We’ll just get the nuclear partners, which is what we want. Then clearly, we could also… Exactly… We could test the effect of depletion of any of those RPLs on the… It’s actually a big experiment. We’ll need to create degron strains, great degron lines for the RPLs, because these cells will not live very long with the knockdowns.

Karla Neugebauer (01:16:19):
We are already uncomfortable about these knockdowns that we conducted in the screen. We are going to want to make auxin-degron strains, and deplete for the shortest amount of time possible. But then we want to do that exact experiment that you suggested. That’s a great experiment.

Antonio Domingues (01:16:36):
I think that’s the last one, and then we’re done.

Karla Neugebauer (01:16:39):
I could say forever. We can stay.

Bede Portz (01:16:42):
Diego, may I please? Antonio, rather, may I please ask one more?

Antonio Domingues (01:16:46):
Yes.

Bede Portz (01:16:47):
I’m super interested in the, I guess, partial miscibility of the Cajal body and the gem. There’s this invagination, but there’s not a mixing, right? Is that a function of they have fundamentally different material properties or surface tensions? Are there some biophysical underpinning this? And also in your CRY2 experiments, have you ever taken the gem protein and the Cajal body protein in the same cell line, both with CRY2, and try to nucleate a fully mixed Cajal body gem thing?

Karla Neugebauer (01:17:32):
I do have some slides that I thought people might want to see. Here’s what we did. We didn’t do what you just suggested. What we did was we took… These are the condensates formed by SMN-Tudor in red. They’re out in the cytoplasm. They’re the red. And in the nucleus–this is a mouse nucleus; that’s why you see the blue blobs, right, centroid bodies. Then in the nucleus, they’re yellow because they’re costained with coilin. The condensates in the nucleus do recruit coilin. But as I mentioned, you can knock down coilin, and you still get condensates. That’s partly why we believe there are other ligands.

Karla Neugebauer (01:18:14):
That’s as close as we got to the experiment that you just suggested. We know they’re in there, but they don’t depend on coilin. We also did a lot of staining for other things, including bizarrely HNRNPU, but we didn’t see staining. I mean, you know what antibody staining is like. It’s a whatever. HNRNPU was that guy that I was very excited about that really increased its binding when we degraded RNA. This is just a stab at trying to find out who could be in there, and so we did see sDMA. We did see some coilin, and we did see some stain with this Y12 antibody.

Karla Neugebauer (01:18:56):
There are a lot of dimethyl… It also suggests that it’s very specific. There are a lot of dimethylarginine containing proteins that we already know about that are not in there, so that can’t account for condensate formation. So, there’s a lot we don’t know. It’s one of those results that raises way more questions than you had before.

Antonio Domingues (01:19:19):
Which is a nice segue-

Karla Neugebauer (01:19:21):
To stopping.

Antonio Domingues (01:19:24):
Because this is the one… Because-

Karla Neugebauer (01:19:26):
Antonio has plans for this evening.

Antonio Domingues (01:19:30):
Yes, because it’s just wonderful talk. Thanks a lot.

Karla Neugebauer (01:19:33):
Thank you.

Antonio Domingues (01:19:33):
It generates a lot more questions. It answers quite a few. Please, Jill.

Jill Bouchard (01:19:39):
Yeah, this has been a fantastic discussion. We just thank you so much for sharing all your work. I think we’re going to need to ask you if we can email you with more questions. Would you be okay if we send or post your email with your video, because it seems like there are a hundred more questions coming?

Karla Neugebauer (01:19:58):
Okay.

Jill Bouchard (01:19:59):
I don’t know a hundred, but-

Karla Neugebauer (01:20:00):
Sure.

Jill Bouchard (01:20:01):
The point is I think we’re fully stimulated with or without coffee today. Thanks to your wonderful talk today.

Karla Neugebauer (01:20:07):
Excellent. Sounds good.

Jill Bouchard (01:20:09):
Thank you so much, Karla.

Karla Neugebauer (01:20:10):
You’re welcome.

Jill Bouchard (01:20:12):
We will have more talks in the future. We have another one coming up in June. So if anybody wants to join that one, come back. Thanks, again, everyone for joining today, and you all have a great day.

Karla Neugebauer (01:20:26):
Bye. Thanks for everyone’s questions. That was great.

Jill Bouchard (01:20:28):
Thanks again, Karla.