VIDEO: Tatjana Trcek on Germ Granule mRNA Self-Assembly

Tatjana Trcek (02:09):
The spatially organized mRNAs are also very important in early development. An example is a Drosophila embryo. So these early embryos are completely transcriptionally quiescent and about 30% of all mRNAs are deposited in this embryo are somehow spatially organized. This really is important because it helps build a body plan of this developing organism.

Tatjana Trcek (02:32):
However, this process can also go wrong as you can imagine and can cause disease. An example is myotonic dystrophy. It’s a genetic disorder in humans that is caused by a massive overexpression of CTG repeats in the 3′ UTR of these mRNAs. When there is aberrant mRNA, is it transcribed instead of being transported out into the cytoplasm, it accumulates into the nuclei and forms inclusions, this RNA aggregate. And these aggregates really are the ones driving the disease. This is an on again function mutation and causes progressive muscle weakening and premature death and of course, myotonic dystrophy 1.

Tatjana Trcek (03:16):
And as [inaudible 00:03:18] have shown, the likely cause of aggregation or assembly of this inclusion is the CTG repeats that are massively over expended where the CTG repeats can base pair with neighboring DMPK mRNAs to cause this aggregate.

Tatjana Trcek (03:38):
So it’s fundamentally important for us all to really know which mRNAs it must spatially organize and which it must not. And when to organize them and where in the cell to organize them, because that can really mean the difference between normal cell growth as well as the disease.

Tatjana Trcek (03:54):
My lab is interested in a particular type of spatial organization. We have mRNAs and this is enrichment of RNAs and RNA granules as you know very well. We know a diverse type of RNA granules, those that are forming in every cell type, such as processing bodies and stress granules. And those that are more special in form in the particular types of cells, such as germ granules that form in germ cells or neuronal granules that form in neurons.

Tatjana Trcek (04:27):
What is common to all of these granules is that they’re all round, they’re non-membrane bound, they’re full of RNA-binding proteins and mRNAs, and it is thought that very important post transcriptional regulation is occurring in these granules and that this post transcriptional regulation is fundamentally important for specific functions that are required by the cell. For example, the germ granules are thought to regulate mRNAs, then regions granules is to control the germ cell fate. Neuronal granules are thought to promote a Neuronal shape and function, processing bodies are thought to be the sites of mRNA decay, and stress granules are thought to be the sites where mRNAs accumulate during stress and are protected and translation silence during stress.

Tatjana Trcek (05:17):
Many of these functions especially in terms of post transcriptional regulation and the mRNA, what happens to these mRNAs, how they arrived to these granules are still largely unknown, though much work has recently been a focus on addressing some of these problems. We now understand that specific features, either sequence features or structural features in the RNA itself are required for enrichment or recruitment of these mRNAs into granules. However, it is largely unknown what happens to the mRNA once it arrives to granules? How are they organized? How are they relating to other RNAs and how they are actually regulated in terms of post transcriptional regulation?

Tatjana Trcek (06:04):
In some of these topics, specifically, how RNAs are organized in these granules I will talk about today. To address these questions, we use the establishment of the germ line in the Drosophila to understand this process. We really like this model system because it first and fundamental steps in the establishment of the germ line of germ cells, really rely on post transcriptional regulation, in fact exclusively initially. The first critical steps involves localization of the subset of mRNAs through the posterior tip of this fertilized egg.

Tatjana Trcek (06:44):
Where the mRNAs accumulate in germ plasm, this is a specialized cytoplasm that accumulates particular proteins called germ plasm proteins. And the germ plasm proteins condense into small granules called germ granules, as shown in green here. And they attract mRNAs. These mRNAs become regulated post transcriptionally in these granules. They’re only translated in these granules that the counterpart that are not enriched in granules are transitionally repressed.

Tatjana Trcek (07:19):
They’re protected from decay and the protein product that are coded by these mRNAs then instruct the establishment of germ cells. So to understand what happens with these mRNAs and how they’re organizing in these granules, we used fluorescent in situ hybridization and super-resolution microscopy. We used commercially available smFISH probes as our small DNA primers that are labeled with a single fluorophore and we used many of them that are aligning on different parts along the mRNA. And this really was helpful because it amplified signal-to-noise ratio, of course increased our detection sensitivity.

Tatjana Trcek (07:58):
We coupled smFISH with a structured illumination microscope. This was a particular version of this type of microscope developed by Hari Shroff at NIH and we’ve been collaborating with him for many years. This microscope provides a resolution that’s about twice better than a commercial confocal microscope and really in combination with smFISH, enables us to see how mRNAs are distributed both in germ plasm as well in the somatic regions of the embryo with very high definition. We looked at one of these mRNAs, nanos I’m showing you here an example and what we found was this RNA outside of the germ plasm here in the somatic region of the embryo, is mostly present as a single mRNA but as soon as it enriches in granules, it forms clusters where these clusters contain multiple mRNAs derived from nanos gene.

Tatjana Trcek (08:53):
On average the number of nanos and RNAs per nanos cluster is about 14 mRNAs. And as I said, we see this type of clustering in granules and single mRNAs being outside for many mRNAs that we have looked at that have a capacity to become enriched in granules. So we wonder how does mRNA clusters belong to different genes relate to each other within the granules? We looked at cyclin B, nanos, pgc and gcl mRNA clusters. We labeled them with smFISH probes, here showing in magenta and we asked how they’re relating to each other but also how they’re relating to the protein content of the granule, labeled here with the core granule protein Vasa tagged with GFP.

Tatjana Trcek (09:37):
And what we found was that they were not all homogenously distributed, instead some mRNA clusters such as those belonging to cyclin B, were located in the center of the granule, those belonging to gcl were located more at the periphery and nanos and pgc mRNA clusters seem to be localizing somewhere in between. We wanted to characterize this colocalization better quantitatively and we employed two measures of colocalization to do this. In the first, we used a spot detection algorithm that measured the nanometer distance between overlapping mRNA clusters and germ granules and in the second, we used Pearson’s Correlation Coefficient Costes approach. In this approach, we measured how the fluorescence intensity of two object that are labeled with spectrally distinct dyes are correlating with each other within the overlap.

Tatjana Trcek (10:35):
PCC Costes is a really great measurement because it’s insensitive to the duration and the frequency of the overlap, so you can do these measurements even in crowded environments such as germ plasm. It ranges from one, either getting perfect colocalization to zero, indicating random colocalization. As the distance between overlapping objects is increasing, PCC Costes is decreasing as shown here, indicating that now as it was the case without measurements, mRNA clusters in granules are now colocalizing with a greater, yet fixed distance relative to each other.

Tatjana Trcek (11:15):
So we use these two measures and implemented them on our mRNA clusters and indeed we found that some mRNA clusters, such as those belonging to cyclin B, reside in the center of the granule. You can see very high PCC Costes for cyclin B and short distance, whereas those belonging to gcl clusters localize much more at the edge. You can see smaller PCC Costes and longer distance and nanos simply just see clusters that are found somewhere in between.

Tatjana Trcek (11:49):
We then use the nanometer distances that we have obtained when we were measuring distances between individual mRNA clusters and clusters and Vasa granules and use triangulation experiments. We’ve done this together with Timothee Lionnet at NYU and we simulated an average three dimensional structure of the mRNA bound germ granule. And what this experiment really showed us, was indeed that these clusters were occupying distinct positions within the granule. These clusters weren’t mixing and that therefore they were homotypic in nature. But what was really, really surprising was that when we looked at core granule proteins, Oskar, Vasa, Aubergine and Tudor, these are the proteins that arrive at the posterior first, make the granule and recruit all other components directly or indirectly, including mRNAs. These proteins in turn were homogeneously mixed. So you could see a lot of white overlap.

Tatjana Trcek (12:53):
So that was perplexing to us. How was is possible that within this unorganized protein environment, you get these mRNA clusters that are so highly spatially organized? And this is what we wanted to pursue and this is the work that hasn’t been published, but it’s now in review and in the final process of revision. It will soon be published. So very early on we’ve noted that mRNAs that can form abundant mRNA clusters, meaning they have many mRNAs per cluster, will reside in the center of the granule, as shown by this curve here. And those mRNAs that form less abundant mRNA clusters will reside at the edge.

Tatjana Trcek (13:36):
For example, cyclin B mRNAs form clusters that are on average contain 30 cyclin B mRNAs, and reside in the center. Gcl clusters on average contains 2.8 gcl mRNAs and reside in the edge and nanos, somewhere in between. So then we thought to deregulate or to change the expression level so there’s mRNAs in the entire embryo. And to thereby change the abundance of these clusters at the posterior pole to see what happens.And we first focused on cyclin B and nanos and we used underexpression experiments and drove down the expression of these RNAs and these embryos now form less abundant cyclin B clusters as you can see here. And less abundant nanos clusters as you can see at the bottom. And this really caused these clusters to move towards the periphery of the granule.

Tatjana Trcek (14:31):
And then converse experiment, we then increased the expression of gcl and when we did that, embryos formed more abundant gcl clusters and now these clusters were pushed toward the center of the granule. So clearly, within this unorganized protein environment, in vivo, these mRNAs have a capacity to self organize and even determine their own position of these clusters within the granules by themselves in an mRNA concentration dependent manner.

Tatjana Trcek (15:04):
This was also very intriguing to us and we wanted to understand it a little bit better. We were intrigued by this ability of mRNAs to self organize and we wanted to understand how. So, the two models that we thought about was that this self organization, self assembly could be mediated by RNA-binding proteins that are deposited on the mRNA in a sequence dependent manner. Or that mRNAs are engaging in the record base pairing through intermolecular trans RNA:RNA interactions.

Tatjana Trcek (15:34):
We wanted to test this. So, in this model homotypic clusters can then arise either through specific protein:protein interactions, specific RNA:RNA interactions or some combination of the two. The way to test this, we thought that if we had a scenario where we expressed two mRNAs that shared RNA features that are required to generate the specificity for homotypic assembly, but were otherwise distinct, then these two mRNAs would populate the same granule and also condense into the same homotypic cluster. But if those features were not required to generate the specificity for homotypic assembly, then these two mRNAs would come to the same granule, but form two distinct clusters.

Tatjana Trcek (16:25):
So we focused again first on nanos and we started with the nanos 3′ UTR because the Lehmenn [inaudible 00:16:33] Lab have previously shown that specificity requirement for the enrichment of this RNA to the germ granules, is encoded in the 3′ UTR of this transcript. So, we thought that maybe features that are driving homotypic assembly of nanos in granules might also be encoded in the 3′ UTR. So we created chimeric mRNAs that were fused with 3′ UTR nanos and were fused to the open reading frame of some foreign transcript ORF belonging to the far right protein. And this also accomodated binding of spectrally distinct probes and differentiation between endogenous nanos and this chimera in the embryo and were driven by some form of promoter and would have a 5′ UTR of some other gene.

Tatjana Trcek (17:26):
And when coexpressed and labeled with smFISH probes, indeed they were colocalizing to the same granule, but they were not coming together to the same homotypic cluster. You can see there’s not a lot of spatial overlap between the chimera and the endogenous gene and even the quantiification itself indicated that the position was low.

Tatjana Trcek (17:47):
So this experiment really indicated that the 3′ UTR features, although required to bring this RNA to the granules are actually completely dispensable for homotypic assembly once this RNA arrives to the granule. So the features are not for this homotypic assembly are not found in the 3′ UTR. We then tested also the promoter, the 5′ UTR and open reading frame either in the combination or alone and found that none of these features were required for homotypic assembly of nanos mRNA within granules.

Tatjana Trcek (18:26):
So this idea that these RNAs are assembling because of specific sequences either through RNA binding proteins, where they have record base pairing, we found not to be true. And in fact, the data that I have just showed you indicates that there is no specific and mappable region, an essential region in nanos RNA required for homotypic assembly in granules and that therefore homotypic assembly of this transcripts is RNA sequence independent.

Tatjana Trcek (19:00):
Again, these results were intriguing to us and suggested that if there should be interactions among these mRNAs in clusters, than they should be non specific or disordered. And they shouldn’t be sequence driven or ordered. We wanted to address that in vivo. First we wanted to make sure that we can actually detect sequence specific ordered interactions in vivo using microscopy.

Tatjana Trcek (19:29):
To do that, we turned our attention to Oskar mRNA. Oskar RNA localizes to the posterior tip of the developing oocyte and work from Anne Ephrussi’s lab at EMBL has shown really, really nicely that this RNA localizes to the posterior tip using first active transport, where this RNA finally engages with the kinesin motor to become transported to the posterior pole. The second Oskar RNA piggybacks on the first one, using a palindromic sequence you can see here in magenta. Where this palindromic sequence is presented within the stem loop that it is embedded within the 3′ UTR of Oskar RNA. And the Ephrussi Lab has shown this using genetic and molecular techniques that interest trans compensatory mutations to really show that these RNAs is engaged in dimerization events using direct base bearing between two palindromic sequences.

Tatjana Trcek (20:29):
And once this Oskar localizes to the posterior pole, they actually also forms clusters there. So we thought that if we use Oskar clusters as the initial readout and we put a single fluorescent FISH probe proximal to this palindromic region… because at the posterior pole, we did this region, Oskar will be engaged in direct sequence specific RNA:RNA interactions and will become spatially constrained within that region. This RNA is going to be bound and will therefore form a very small radius of the RNA cluster as reported by this single probe. But if on the other hand, we now put a single probe somewhere else on the Oskar RNA that is not engaged in sequence specific trans RNA:RNA interactions, then within that region these RNAs will not become spatially constrained and we will be able to measure a much bigger radius of the mRNA cluster.

Tatjana Trcek (21:32):
This configuration then in vivo will suggest to us that indeed, this particular RNA is engaged within sequence specific ordered trans mRNA:mRNA interactions within the 3′ UTR of this transcript. Finally, if regardless of where we put the probe, 5′, 3′ UTR or middle, we detect more or less the same radius of the RNA cluster and this configuration is going to indicate to us that this RNA, in vivo, is engaged in non-specific disordered trans RNA:RNA interactions.

Tatjana Trcek (22:10):
So to do this, we used STORM microscopy and we collaborated with Eli Rothenberg’s lab at NYU and his postdoc Yandong and we labeled an RNA such as nanos with a single FISH probe that was conjugated with a photoswitchable dye such as Alexa 647. We observe very, very long and beautiful photoswitching events as shown here on this blinking movie. We then reconstructed the image, extracted the data and fit the data to an autocorrelation function as you can see here on this graph.

Tatjana Trcek (22:47):
By doing this we were able to extract information of the radius of the RNA cluster as reported by that probe. And we first looked at Oskar RNA and we hybridize probes one by one along the entire length of the Oskar mRNA and as soon as we hit the palindromic sequence, we recorded a reduction in the cluster radius. About 40% reduction was statistically significant, which indicated that this RNA in vivo, within the 3′ UTR, was engaged in sequence specific or trans RNA:RNA interactions.

Tatjana Trcek (23:25):
This was beautiful because it was aligned very well with the molecular and genetic data reported by Anne Ephrussi’s lab. And I would like to point out that the radius of the cluster is not indicative of the length of the interaction itself, it just shows where a particular region of the mRNA is most likely to be found. It’s a probability measurement. But when we turn our attention to nanos RNA that is clustering in germ granules, we found no such constriction, regardless of where we put the probe. What this result indicated to us was that in vivo, in germ granules in these clusters, nanos is engaged in non-specific disordered trans RNA:RNA interactions, despite it’s very high concentration.

Tatjana Trcek (24:17):
The data that I showed you really suggest the following model and that is that upon their enrichment, two germ granules, mRNAs form homotypic clusters and that they do so in a sequence independent manner, but in a highly transport specific manner, in such a way that these RNAs are able to distinguish between foreign RNA such as nanos and gcl and as well as near cognate RNAs such as nanos and its derivatives.

Tatjana Trcek (24:46):
We of course do not know what generates the specificity for homotypic assembly, but given that small perturbations to the transcripts are enough to cause the mixing of a chimeric mRNA from the endogenous RNA, this result suggests that it’s probably the entire property of the mRNP that’s really important. Not just sequence, not just single structure or a protein bound to it. So in this holistic approach, and this is something that we are trying to understand now, we would like to understand how the proteins are bound to the RNA contribute. Proteins that are directly bound as well as perhaps ribosomes. How the structure has contributed, how the modifications are contributing towards generating the specificity for homotypic assembly.

Tatjana Trcek (25:37):
What we also would like to understand is what happens to these mRNAs after they become enriched in this very clustered and highly concentrated condensates? What was fascinating to us was the measurement that outside of the posterior pole in the sematic regions of the embryo, mRNAs such as nanos have less of a concentration. We measured about 3.2 nanomolar. But once they come to the posterior pole and form these clusters, they become highly, highly concentrated. Almost 5,000 fold increase in the concentration within these clusters.

Tatjana Trcek (26:12):
There’s plenty of opportunity for the mRNAs to become engaged possibly into promiscuous and potentially the deleterious interactions that would not be beneficial for the embryo. But this is not what we see actually. This embryo is perfectly fecund and we would like to understand how is it possible that in cases such as myotonic dystrophy, when you have this sequence dependent interactions that form aggregates and cause disease, on the other hand we have sequence independent, nanos clusters that are also highly concentrated but promote fertility? What it is… the difference between the two of them besides sequence dependence and sequence independence, and how are these potential conflicts that could give rise to disease result in germ granules and homotypic clusters that form there.

Tatjana Trcek (27:10):
Finally, I would like to point out that we see this type of homotypic assembly for every mRNA that we have looked at that is capable of enriching in germ granules and importantly other people have observed such assemblies as well. For example mRNAs that localized germ granules in zebra fish as well as mRNAs that localized stress granules and processing bodies in U-2 OS cells.

Tatjana Trcek (27:36):
So it seems that this is a common feature of the mRNA that is revealed at sufficient concentration. Possibly, when they become enriched in this, type of condensates. We would like to understand, why RNAs do this? Is there a function to this homotypic assembly? We would like to think about this homotypic assembly as post transcriptional factories. As I said, translation, regulation of RNA turnover, prolonged localization, even perhaps buffering of gene expression, noise occurs in these granules and we would like to think that maybe these homotypic clusters have something to do with it.

Tatjana Trcek (28:18):
Just as the concentration of this RNA increases, once RNAs localize to these granules, so does the concentration of important regulators, effective molecules that these RNAs bring with themselves that could then possibly regulate their post transcription after the RNAs enrich in granules.

Tatjana Trcek (28:38):
This is something we’re pursuing in the lab as well. In the end I would like to acknowledge my lab for really being courageous in pursuing these questions. We are the beginning we just started our lab but we are trying our best to tackle these difficult questions. Of course, I would like to thank you, Dewpoint Therapeutics, for inviting us to participate in this seminar and to discuss these discoveries with you. If you have any questions I would really like to invite you to ask us and email us and we are looking forward to interacting with you more. Thank you.

Mark (29:18):
Thanks. It’s wonderful.

Tatjana Trcek (29:20):
Thank you.

Mark (29:21):
As you know, there’s an approach in Germany, they don’t clap, they bang on the table.

Tatjana Trcek (29:29):
Oh.

Mark (29:30):
We’ve gotten used to doing that instead of clapping. Thank you. Questions? I bet there’s a bunch of questions.

John (29:40):
Hi.

Tatjana Trcek (29:40):
Hello.

John (29:42):
Great talk.

Tatjana Trcek (29:43):
Thank you.

John (29:44):
I want to ask a technical question about the single molecule RNA FISH. On that slide where you had all the different dots shown and you had a measure of the number of RNAs per dot, how did you estimate the actual number of RNAs in each spot? Was that just based on intensity?

Tatjana Trcek (30:00):
Yes. So we are using an algorithm that was developed by Timothee Lionnet. It’s an algorithm that uses the intensity of single mRNA as a proxy and then we calibrate and normalize all other spots relative to that. Yeah. The nice thing about FISH, or maybe the bad part too, it’s really great at lower concentrations. We find that about 3.5 nanomolar is really the upper limit. So we can’t really quantify anything… we would never be able to quantify beta actin, for example.

Tatjana Trcek (30:44):
But we are very sensitive to lower concentrations and we are very accurate. But we can calibrate bigger spots nicely knowing what the intensity of a single one is.

John (30:57):
Right. And the reason you can’t go above 3.5 is just because it gets so… there’s signal everywhere, so you can’t distinguish anything?

Tatjana Trcek (31:03):
Yeah. It becomes too dense and you can’t resolve them very nicely. So what really helps is structured illumination microscopy. First, it gives better resolution and then it gives better signal-to-noise so you can differentiate them better. We see a very nice correlation between single molecule FISH data and RNA-seq in terms of RPKM values, so we know that we are doing it well. But it breaks down at higher concentrations.

John (31:34):
Cool. Yeah, because one thing that we’re interested in here is doing the FISH on a bunch of different potentially interesting condensate targets and one things I’ve always thought about is, since FISH is always a puncta, right? The signal is always a puncta, so how do you decide whether this cluster of puncta is maybe in a condensate or not? So that helps, that helps a lot.

Tatjana Trcek (31:54):
Right. So first, this is a real classical problem, because in the end, even if you don’t add any probabilistic puncta, that just the nature of biology and there’s a background. First, it really, really helps if you have many mRNA and many probes. So I can send you the protocol, at least for what we’ve written… I’ve written a protocol for Drosophila and then also for yeast. It’s more or less the same and it’s always the same problems. It helps to start off with something that you can change the concentration of, so that you can go to a higher or lower regime and you can see how spot change. Because sometimes, we all imagine these beautiful spots and then you do it or experiment and you just don’t see that and you’re confused.

John (32:50):
Yeah, that will be great, thank you.

Tatjana Trcek (32:50):
Right. So I would say I would start with something that is not as expressed as beta actin if that’s a mammalian cell, if that’s a yeast cell, I would start with… I can give you some examples and if you’re working with something I can give you examples as well. Then you would also do classically, you can do a knockdown to see that the signal disappears, you could do a two color experiment where you mix probes that are red and green. But really, starting off with many probes is better. And choice of fluorophore in my hands, the fluorophores that we use and are mostly reliable, really, really nice are [CAL Fluor® Red Inaudible 00:33:40] 590, we’re using Stellaris probes, they’re commercially available, they’re very easy. When I was working with yeast, it was Cy3 that was really reliable. Far red probes are also okay, but they tend to bleach and 488 is beautiful, too.

John (34:00):
Okay, awesome thanks so much. Sounds like you’ll be a very good person to talk to if we have technical problems.

Tatjana Trcek (34:04):
Oh yeah, yeah. It’s not that difficult, it sounds crazy but it’s true. You need to learn to see. It’s a little bit, at first, biased. Because you’re hoping to see those published results and then you look at your gene and it’s just not like that and maybe it’s just under expressed or maybe it’s so highly expressed that everything becomes very homogenous so it looks like diffused background.

John (34:33):
Yep. Makes sense. Awesome thank you.

Tatjana Trcek (34:36):
Sure.

Mark (34:38):
Emily, did you have a question?

Emily (34:40):
Yes. Yes I have a question. This is beautiful talk. I really enjoyed all of these diagrams and also the images of all these fluorescent spots. In graduate school, I used to do a lot of [in situ inaudible 00:34:53] on the [inaudible 00:34:56] regeneration development on the [inaudible 00:34:59], where the mouse embryo was taken out and we would do in situ hybridization and see all these morphogenesis environment where all this RNA was… during development, expressed a different pattern formation. Also, the [inaudible 00:35:19] has the [inaudible 00:35:20] region lighten up with different mRNA, so it’s very exciting to see all this in an embryo.

Tatjana Trcek (35:27):
Mm-hmm (affirmative). Yeah.

Emily (35:29):
So I have a question that, I think I missed the point. Where you showed a different probe, it had a different effect on the same mRNA. Where some of them close to the [inaudible 00:35:46], it has more aggregation, but the other region, it’s a little more diffused.

Tatjana Trcek (35:53):
Right.

Emily (35:55):
I’m wondering, what is that?

Tatjana Trcek (35:57):
So, we observe clustering for all this mRNAs so when we did genetic experiments and we were swapping regions, we found that not a single region was required to generate specificity. In other words, there’s no element that would bring two mRNAs together in a specific manner. So that indicated to us that if there are interactions among these mRNAs, and this is a highly concentrated environment, it shouldn’t be sequence specific. It should not be something where the mRNAs, even though they’re highly concentrated, should be extremely highly packed. We wanted to address this and we looked at an RNA where we knew that the RNA was engaged with a particular part of that mRNA in direct base bearing with another RNA originating from the same gene. That was Oskar.

Tatjana Trcek (36:57):
So it has this little palindromic sequence that is able to dimerize and only within that region, that little stem loop. And then when we looked in the embryo and we looked at what was the radius of rounded palindromic sequence, because these RNAs in the cluster were engaged in dimerization, they were much more closely aligned to each other than with any other part.

Emily (37:24):
Right.

Tatjana Trcek (37:25):
Even though there could be interactions among these clustered RNAs in a highly crowded environment, those infectious that are occurring outside of the palindromic region, they don’t generate as much packing ability as a single sequence specific region, a palindrome. It indicated to us that: A, outside of the palindromic region for Oskar, that mRNA is likely… I’m not going to say flopping around, but the interactions are much more transient labile and that mRNA might be a lot more fluid outside of the palindromic region than within the palindrome, which constrains the mRNA.

Tatjana Trcek (38:18):
Then when we looked at our RNAs, we didn’t see any such contrition. In fact, the radius was more or less the same along the entire RNA as we observed to Oskar outside of palindromic region. Which indicated to us that along the entire RNA in the homotypic cluster in granules, we don’t observe sequence specific base bearing. The interactions are there for disorder and that the entire content of mRNAs in the cluster is likely diffusible, mobile. There’s no constriction and that was the relevance.

Emily (38:58):
Really? Thank you. I had a really unrelated question. When RNA was transcribed, it was transcribed as a pre mRNA where all the intron was also attached and then you processing to cleave off the intron region. I was wondering if the clustering happens when the intron is still attached without cleave it off, because a lot of intron has repetitive sequence where maybe you have the all GGG, CCC, that’s kind of homotypic every day before it was processed, but sometimes if you use your probe only looking at the translated region, you detect it but you did not detect the intron which is not cleaved yet.

Emily (39:54):
I was wondering if we know in the germ granule, all this mRNA is already processed so all the intron was already cleaved off.

Tatjana Trcek (40:06):
That’s a really good question, we haven’t thought about that. I’ll answer two ways. The general overall RNA-seq data in the embryo suggests that all these RNAs were processed. But only 4% of the RNA actually localizes in these granules of the total. 4% of nanos. So, it could be that we’re not detecting the 4% or that the 4% that localized have the spots, so we don’t know. That’s a good… We should see. But I can tell you that all the chimeras besides the last one where it was the full length nanos mRNA, together with intron, so it did an entire gene region. So all other chimeras actually didn’t have introns. They nevertheless localized and clustered. So, I don’t think so, but we actually have not verified that this 4% that do localize, don’t have intronic regions.

Emily (41:09):
I see. Thank you.

Emily (41:12):
I have a last one.

Mark (41:12):
Okay.

Emily (41:15):
This is more theoretical, do you think this is a storage depot for them? Or how they were moving out of it to get translated into functional protein?

Tatjana Trcek (41:24):
This is all indirect evidence. We don’t actually think that RNAs exit these granules to become translated. The reason for that is, we’re developing tools to address exactly that and some other questions regarding translation and homotypic clustering, but nanos RNA has a translational control element that represses this RNA in the somatic regions of the embryo. It binds this particular protein called Smaug. It requires Oskar, which is almost exclusively found in granules, to relive the repression imposed by Smaug. So I don’t see how it can go in and out so that it would be enriched, but then if it needs to be translated, it needs to exit because it will just be repressed again, and actually, would diffuse away because the embryo goes [inaudible 00:42:25] and it’s about 500 microns across the volume of germ granules is 0.01% of the entire volume, and if I believe that if the RNA cluster would exit, it would just diffuse away.

Tatjana Trcek (42:38):
We don’t see any loss of mRNA over time in this embryo. Also, the gradient of nanos protein is originating from the posterior and goes towards the anterior, so we know that it’s produced locally, but nobody has really formally tested it. I think it’s really indirect evidence that suggests that translation is happening in granules.

Mark (43:02):
That’s great.

Tatjana Trcek (43:04):
Thank you.

Mark (43:05):
Alicia, you had a question, I think?

Alicia (43:09):
Yes, hi. I was just wondering if you have any structural evidence to support the model that these RNAs are disordered.

Tatjana Trcek (43:22):
We don’t know anything about the structure of the RNA itself. Zero. We would like to have that data. It’s a little bit more complicated just because of the availability of the material, that’s number one. 4% of RNA enriches in granules that are only found at the posterior end and account for 0.01% of volume. And although these granules have gel like properties, they also have liquid like properties, so purification is going to mess them up a little bit. We would really like to know what is the structure of the RNA itself in granules? We don’t know that. All we know is that the interactions among them are not anything that is ordered or sequence specific and that’s what we’re calling disordered.

Alicia (44:21):
Is there a way that you can do and NMR or any other method in order to show that?

Tatjana Trcek (44:29):
So the genetic evidence, when we were pairing mRNAs along that shared regions and we saw the chimera and the endogenous mRNA never came together even though they shared exact regions, this chimera recapitulated all the essential features of the endogenous RNA, so we thought at one point, if we add all the mRNA sequences back to the chimera, we should recapitulate it until our RNA and they should come together in a single stretch or combination of these stretches is required to generate specificity. We didn’t see that. So that’s why we’re thinking it’s the entire mRNP and all the characteristics of that mRNP, they are important.

Tatjana Trcek (45:19):
And then with STORM, we saw the same thing. Now, we don’t see any particular constriction, we are not saying that there’s no interactions, we just say that there is nothing that is sequence based. That is specific, right? And that’s why we’re saying that these interactions, should they occur, are disordered. They’re not sequence specific ordered.

Ann (45:47):
Mark, I have a question.

Mark (45:49):
Great.

Ann (45:49):
I was just searching for an analogy here…

Tatjana Trcek (45:54):
Oh.

Mark (46:02):
The same thing that was happening to you yesterday, Ann. You keeps muting you.

Ann (46:07):
Yeah. So in music, you can listen to a piece of music and it has a feel, right? This is a baroque piece, or this is Beethoven, this is Schubert, this is something entirely different, Aaron Copland, right? If you’re really familiar with music, you can just listen to a little excerpt and… My daughter and I would play this game of guessing the composer or guessing the piece, but the notes are all the same. And yet there’s this feel that is different. Somehow I’m wondering if, in a way the cell is able to listen to music.

Ann (46:48):
This mRNA has a different sound. The whole, even though the notes, the nucleotide are all the same, it’s not that. It’s something bigger. Does that make sense?

Tatjana Trcek (47:03):
We, especially when the paper was in review, we were really trying to… It is a bit confusing, right? We haven’t shown what is required, we have just shown that something is not and this is primary sequence. If primary sequence, either by forming a secondary structure that would be conserved. We are not saying that it’s not important, we’re just saying that as soon as you modify this RNA a bit, and we modified it either by inserting an EGFP open reading frame in one, so it was the entire length of the mRNA with the introns and the promoter and everything. So either we inserted the EGFP or we inserted m, it became different.

Ann (47:47):
Right.

Tatjana Trcek (47:48):
Even though it is still the same mRNA, it recapitulated all other features that we could characterize for that mRNA translational onsets, transitional repression, localization, clustering, and yet, it was behaving differently.

Ann (48:05):
Yep.

Tatjana Trcek (48:06):
That’s why I was saying it’s not as simple as one defining or mappable structure. And on the other hand I think that what makes it… In terms of your analogy and music, you can listen to a little piece and say, “Oh that’s Coltrane” and in our case you’ll look at the 3′ UTR and say, “Oh, that needs to be bound to germ granules.” But once it gets there, there’s all this other stuff that makes this RNA nevertheless behave differently. So there’s little nuances that then matter.

Tatjana Trcek (48:38):
What I think is really interesting and important, it’s because of this high specificity, perhaps, that you don’t really see often clustering outside. Even though these RNAs share features with many, many, many other RNAs that should, in principle, come together. I will talk about exactly that. This is all very preliminary, as I said I’ve just started, but we do have sequences in these RNAs, especially the 3′ UTR on this transcript that made them specifically bind to each other.

Ann (49:17):
Yeah, you have that palindrome interaction, so that can override everything. But I guess what I would say is that if you have a piece that’s Schubert overall and then you stick in a little piece of Mozart somewhere in there, it doesn’t change the Schubert overall, so that’s what you’re doing. We just don’t have a way, but the cell must have a way of sampling this is EGFR, this is [inaudible 00:49:42], and it’s not any one small little piece of the RNA. We’re reaching for a different metaphor, but are reaching for a different measurement. It’s kind of like, being color blind and also now being able to see, “Oh my God, this is green.” We’re only looking at black and white. But black and white just looking at the tiny tiny motifs in the RNA, it’s clearly not that, it’s something else.

Tatjana Trcek (50:06):
Right, so we call it solubility. I know it’s common in the phase separation. But if they come together, they’re soluble. If they don’t, they’re not. And what makes them soluble or not soluble is some property of the entire molecule.

Ann (50:24):
Yeah, but to me that’s a very pragmatic description.

Tatjana Trcek (50:29):
It is.

Ann (50:32):
I think it’s probably more poetic than that when, as you do my experiments, it’s going to be something bigger than that. It’s super cool. Beautiful, beautiful talk. Thank you

Tatjana Trcek (50:41):
Thank you. Thank you so much.

Mark (50:43):
It really was. Obviously we could keep asking questions for another hour, but we’re out of time. But maybe if it’s okay with you if we have other questions, we could just follow up with you?

Tatjana Trcek (50:53):
Sure, and also we love to hear your perspective on this. As I said, we’re a young lab. We like questions, we’re trying to see how other people think about our problem and sometimes we’re maybe a little bit too mired in what we do and we certainly… If you have questions just email the questions to us. We’re excited that we can participate. Thank you so much.

Mark (51:23):
Oh that’s great. So thanks again and thanks for everybody on the call as well for joining and for the great questions.

Tatjana Trcek (51:28):
Thank you. Bye.

Mark (51:31):
So everybody have a great day.

Emily (51:32):
Bye.

Mark (51:33):
Yes, thank you, you were awesome.