HomeArticlesRuth Lehmann (NYU / HHMI) 1: Germ Cell Development
Ruth Lehmann (NYU / HHMI) 1: Germ Cell Development
September 14, 2019
Hello. My name is Ruth Lehmann, and I am a professor at NYU Medical School, and I’m also an investigator of the Howard Hughes Medical Institute. Today, I’m going to tell you about germ cells. When egg and sperm meet, they make a fertilized egg cell. And at that point, a really important decision has to take place. And that is the decision between eternal life or death, between germ cells and soma. At the very early stages of embryogenesis, germ cells are set aside, and they have the potential to give rise to a whole nother organism. All the somatic cells, which give rise to our body, will eventually die. So, germ cells, while not important for the survival of the organism, they are absolutely essential for the survival of the species. And they are the site where evolution takes place as well. So, what I want to tell you about is how germ cells are really in… associated with germ granules. And how these germ granules form. And how they’re essential for the function of germ cells. I will also tell you, in the second part, of how we actually achieve this dichotomy of the germ line and the soma, by setting them aside, and also by inhibiting each other. And finally, in the third part, I will talk about mitochondria, which are only transmitted through the female germ line. So, the term germ plasm was initially termed by August Weismann. What he was referring to was that only the germ cells carried the potential for inheritance, and the elements of inheritance. He was sort of referring more to chromosomes, but we’re left with the term. And also, he was studying organisms which have the classical germ plasm determinants. And so he mixed up what’s in the cytoplasm and what’s in the DNA. Today, we know that there are two parts of how germ cells are set aside, two mechanisms. One mechanism is the germ plasm mechanism, which is a mechanism by which maternally inherited proteins are set aside at a certain location within the egg or early embryo, and those contain the instructive information which then will lead to the differentiation of the cells which inherit that germ plasm into germ cells. The other mechanism, which is actually much more common and used by more animals, is an inductive mechanism. And that mechanism is actually not so different from all the mechanisms that are used to specify the other cells in the body. And in this case, signals from the extraembryonic ectoderm are signaling to the embryo proper to determine germ cells. And so, if we just focus on the germ plasm mode of inheritance, there are a number of organisms which inherit germ plasm and go… follow this mode. And those are, in insects, Drosophila melanogaster; in nematode, Caenorhabditis elegans; and the frog, Xenopus laevis. And in all of these cases, maternal products are deposited in the egg. And those maternal products will determine where the germ cells form. So, these are classical determinants of a developmental fate. And the proof for that kind of determinative ability of the cytoplasm was shown by transplantation experiments. And so, if you use the cytoplasm of an embryo where this germ plasm had been deposited at the posterior pole during oogenesis, which normally would give rise to the germ cells, later on, at the posterior pole, here in the Drosophila embryo… if now, instead, that germ plasm was transplanted — and these were experiments done by llmensee and Mahowald — the result was that this germ plasm then was able to induce cells that have the germ cell fate. So, this clearly established the germ plasm as the determinant for the germ cell fate. However, as I told you, there is another mode. And this may actually be the more common mode of how germ cells are initially set aside. And in this case, just like most other cells in the body, the germ cells are set aside by an inductive event, where two cells signal to each other, and those cell signals then specify fates. And in this case, in the mouse embryo, but similarly in the human embryo and in many other embryos, the extraembryonic ectoderm sends a signal — the bone marrow promoting factor 4 and 8 — to the epiblast, which is the embryo proper. And then the cells which are just at the border, at the interface, they will become primordial germ cells. So, the way it was determined… that in the case of induction there isn’t a determinant which has been deposited in the egg and thereby tells us where the germ cells form… the following experiment was carried out. Cells in the epiblast at… very close to the extraembryonic ectoderm will make mesoderm or primordial germ cells, while the cells at the bottom of the epiblast will give rise to the neural ectoderm. And so, if we now transplant cells which would normally give rise to the neural ectoderm to the new position of where they get the influence from the extraembryonic ectoderm, these cells then give rise to mesoderm and PGCs. And that tells us that there is an inductive event, and that there wasn’t something prelocalized in the egg which told the cells which fate to become. Despite these differences among the way germ cells are specified… and this has often been taken as… perhaps, that germ cells are specified in very different ways in different organisms, and so maybe their nature is very separate and the ways how they act are very different. And indeed this idea was also supported by the fact that there isn’t a master regulator transcription factor for germ cells. So, for muscle cells, for example, in every organism studied, there is a MyoD-like transcription factor which determines muscle cells. That is lacking in germ cells. However, there is something that all germ cells share, and that is germ granules. And they are the hallmarks of all germ… germ cells, irrespective of whether they’re formed by the inductive mechanism or by the germ plasm mechanism, the determinant mechanism. And here are just a few examples. So, here are the germ granules in Drosophila, and I will tell you much more about those. And then you also see nuage in mouse sperm, and the Balbiani body in mouse oocytes. And those are specific granules which are only found in germ cells. And what is also important, and why I would like to refer to the germ granules as a hallmark, is that there’s a lot of conservation between these granules in germ cells. And so that suggests, actually, that there is a common origin, and that there is actually something very similar between germ cells. And so, one is these nuclear… ribo-nuclear particles, which I already mentioned, and I will tell you much more about those. But then there’s also the aspect that germ cells are really… a very, very important function for germ cells is RNA regulation, so how RNAs are localized, translated, the stability, and the Pi RNA processing, which is an important mechanism which occurs in the germ line to defend off transposable elements. And then finally — and that’s maybe the most convincing argument for saying that is something common about among all germ cells — are the RNA regulatory proteins which are conserved. For example, the Vasa protein, which is an ATP-dependent helicase, has been used broadly to identify germ cells in pretty much every species. Other RNA binding proteins, or RNA binding protein interacting proteins like Tudor and the Argonaute protein family, are similarly conserved and are expressed throughout germ cells in all species that are studied. And so, I will be focusing… and this is really the work that’s going on in my lab, is in the germ line of Drosophila. And so, I want to tell you something, first, about the germ line life cycle, and then focus on certain aspects of that life cycle. And so, let’s look at the early embryo, here, which develops within 24 hours into a larva. But when it is laid, it already has a lot of information because of the aspects of the… that the mother provided, because of the proteins that the mother has provided. And so, this is a… pretty much a freshly laid egg. And what you can see… it is… we’re seeing, here, RNA being localized at the posterior pole. And you can see the germ plasm, there, very clearly. This germ plasm then will give rise to the germ cells, only after about two hours, and I will tell you more about this later. Then these germ cells will migrate to the somatic part of the gonad. And there, these primordial germ cells will start to differentiate into male or female germ line. And in the male and in the female in Drosophila, they will associate with a somatic niche, and will continue to make egg and sperm throughout the life of the fly. Also, with this lifecycle, what’s important is that mitochondria are passed on through the germ line. Indeed, they are enriched in the germ plasm and then are carried on through the female germ line alone. So now, in this first part of telling you more about specific research on germ granules, I want to tell you about a really exciting new way of how to look at germ granules. And that is to understand that they are membranous granules, which we knew for a long time from EM analysis. But now we have a better idea how they could be… physically come together, because a granule without membranes is kind of hard to imagine how it could be organized. And so we’re starting to understand, and we started to inquisite, the organization of these granules. And then we hope, by understanding the organization better, we will actually understand their function better, which is really in RNA regulation. As I said, a hallmark of germ cells. So, I will tell you about three parts of germ granules. I will first go back a little bit in history, and tell you about the genetics and assembly of the germ granules. I will then tell you about the biophysical properties which determines that the germ granules are indeed membranous granules. And then I will tell you how the RNA is organized within those granules. So first, I already showed you this image of an embryo. This is actually an in situ hybridization for the nanos RNA. And you can see the germ plasm where the RNA is enriched. And then you can also see RNA, actually, throughout the embryo. You can then see the germ plasm, here, in a… more highlighted. And you can actually kind of gauge that there’s a granular… there’s spaces and… and… and more dense fluorescence. And here’s an EM picture, then, of a single granule. And that is the cytoplasmic germ granule, a hallmark for the fly germ plasm. And what is in these granules? So, many experiments, looking at localization and immunoprecipitation, have led to identify many proteins which are in these granules. And as you can see when you look at this list of kinds or types of genes which are found within the germ plasm and in the germ granules, they’re all RNA regulatory genes. But what is specific for these granules? Because there are a lot of these membranous granules also in other cells than germ cells, and they have many of these same components. But what’s specific are these germ granule assembly proteins and the Pi RNA-associated proteins. And the Pi RNA-associated proteins, again, are required for transposable element control, and I will not be able to tell you more about this at this point, but I will tell you a lot about the germ granule assembly now. So, first of all, again, when we look just at this picture, where we see sort of blank and filled and less filled spots, we can ask… the main components of the germ plasm in the fly, which are really the initiators of granule formation, is the gene Oskar, which then recruits the protein Vasa. And we can ask the question of how much Oskar and how much Vasa are in the granules, or besides, in those blank spots, because that gives us an idea of how the germ plasm is really structured. Do we… can be focus on the granule, or should we focus more on the whole plasm? And so you… what you can see is that the predominant amount of the proteins is in the granules, and only a very small part is actually in the interspace. And that’s not so different from what’s in the rest of the embryo, which will give rise to the soma with its deadly fate. So now, let’s look at the genetics and assembly. The critical gene here is Oskar, and I have to tell you a story about Oskar because when I was a graduate student I identified Oskar, and I named it according to a character in Gunter Grass’ book, The Tin Drum. And you may want to read the book to figure out why I named Oskar Oskar. But the clue here is that it has a patterning defect. They’re really short, these Oskar embryos. They come from homozygous mutant mothers. And so the first was their polarity effect. But what’s really important about Oskar is that there are no germ cells. No germ cells are formed in these Oskar mutant embryos. And when I say Oskar mutant embryos, I mean, actually, embryos which came from mothers that were not making Oskar protein. Remember that there’s this germ plasm which is made during oogenesis. So, interestingly, Oskar is quite an… quite a novel protein, but it is very important as an assembly factor for the granules. And it is possible that these kinds of proteins, which are structural, may be existing in different organisms. And that they may not always be easily recognizable on the basis of their protein composition, of their amino acid composition, because of the nature of how membranous granules are composed. And I will come back to that again. So, what do we know about Oskar? We know a lot about Oskar. After I characterized and identified Oskar initially, a postdoc then in my lab, Anne Ephrussi, really took on to understand many of the aspects of Oskar regulation. And she has, now, her own group at EMBL for many years. So, Oskar effects… is regulated by splicing. It’s RNA is transported during oogenesis into the oocyte. Then, as it gets localized to the posterior pole of the oocyte — and this is critical — it is translated. And when it is translated, then it recruits all these other components, and makes the germ plasm. And then something really peculiar happens later in oogenesis, and that is that all the contents of the nurse cells, which are these feeder cells to the oocyte, they are dumped into the oocyte, and then the oocyte does this kind of washing machine swirling. And so, many other components, and in particular the RNAs, which are components of the germ plasm, get stuck to the germ plasm at that point. And we’ll talk about that later. And that is when the germ plasm effectors are entrapped and localized. So, the Oskar structure has been characterized, both in my lab, in collaboration… collaboration with Rui-Ming Xu, and then also in Anne Ephrussi’s lab, and we know now that there is a very nice structure of Oskar, where it really…we know… also how it interacts with the Vasa protein, and that that is a direct interaction. And then this complex, so a dimer of Oskar, a dimer of Vasa… with Vasa, they recruit other components. And one other complex is the complex between Tudor and Aubergine, another RNA binding protein. And Aubergine has an important role in the Pi RNA pathway. And so these are… these proteins are all assembled together at the posterior pole, upon the localization of Oskar. What I will be focusing on a little bit are these wriggly lines in here, because those are unstructured domains, which with structural biology you can actually not determine. And so we don’t know how they actually behave. And we will think about this more when we’re thinking about how these membranous granules form. But first, why do I say Oskar is the major gene which can assemble the whole germ plasm? And that’s from this experiment, which Anne did a long time ago, again, when she was in my lab. And what we did there… we decided to say, okay, if Oskar has all of this power, let’s put it at a new location and see if, at that new location, it can actually assemble the germ plasm. And so what we did is since Oskar is localized through its 3′ UTR to the posterior pole, we used the localization signal from the bicoid gene, which normally localizes to the anterior pole. And so we just put that on Oskar protein, and we were really lucky, because Oskar got localized, the RNA got localized to the anterior pole, and the protein got translated at the anterior pole. And Oskar did it all. It assembled all the other components. And that led to making an embryo which had, now, two posterior ends. And this comes back to a loss of polarity in the mutants. But most important for what we’re talking about, it had germ cells at the anterior pole, as well as the posterior pole, which is the normal localization. So, this established Oskar as kind of the equivalent of the germ plasm. The interesting part is it would have been very, very hard to find anything like Oskar by biochemical methods. It really needed the genetics to identify the genes which were components of the germ plasm, and then to do genetic experiments, which really told us how this whole system is put together, and how the different proteins are starting to interact with each other. So, we have these proteins which are assembled, and obviously there are many more. I’m just showing a few of the proteins because we know these are sort of the central ones, and these are also the germ plasm- and germ line-specific proteins. The major role, then, is, of these proteins, to assemble what we call effector RNAs. And there about 200 RNAs in the gen… a very generous measurement, that are specifically enriched at the posterior pole, because of the action of these proteins. And these effector RNAs are really what is giving us the function of the germ plasm. Because they are important for germ cell formation and specification. And I will tell you more about this in the second part of the lecture. They are important for germ cell migration. And they are important for transposable element defense. So, the proteins provide a structure, we think, on which these RNAs assemble. And then on which these RNAs can function, and that means translation. And so… and now, I want to turn to tell you more about the biophysical properties of the granules as membraneless granules. And for this I again will step back a little bit, and give you some background of what now is known about these membranous granules and how they are formed. And so, as I mentioned before, there are actually a lot of these kinds of membranous granules. And so, I remember when I was a starting biology student. I sort of thought of the cytoplasm as a liquid. Then I realized it was much more dense, so then it was like a gummy bear. And now we know there are all these structures in there that can form. And so it becomes more like… I don’t know… a chocolate chip cookie or something. So… so, the… the cytoplasm has much more structure than what we thought before. And some of these granules can form in different conditions. So, for example there are the stress granules, which form only under stress situations. Some of the granules are clearly involved in RNA regulation, like the neuronal granules. And then there are also some of these granules which are in the nucleus. The nucleolus, for example, is a very well-known granule that is found in the nucleus. And then here is an example from C. elegans, and that is the germ granules of C. elegans. And they’re also in the cytoplasm. And so the idea is that these membranous granules initially form by demixing from a solution. And so if you imagine you have all these proteins in a solution, and now you increase the concentration, they can demix. And they can demix, for example, when the temperature changes or the pH changes or other conditions. And so, now, the concentration of a particular molecule is much higher in this condensate than in the outside environment. And the idea is that what happens is as these liquids condensating in the granule… that they first, in the granule, can have even the appearance of being quite liquid, which means that they are… can move in and out freely into the condensate. But they can also attain more structured organizations, where they then become more solid and more organized. And perhaps even rigid, where they are no longer able to dissolve. And that could be an explanation for the amyloid bodies that we find in Alzheimer’s patients. So, what are the conditions that favor or influence granule formation? So, I already told you, high concentration, low temperature, low salt, also crowding agents, and all kinds of protein modifications, which often are called, also, multivalency of these… of these proteins. But there’s also specific proteins which are more likely to form these kinds of condensates. And a lot of this is figured out in vitro. Of course, I will be telling you about our experiments in the embryo, trying to understand whether these biophysical properties are also true for our germ granule components. But what is important is that they often have unstructured domains, disordered regions; low complexity domains, that means the same amino acid being repeated many times; and in many cases they have RNA binding domains, or the ability to bind RNA, and often it is that ability to bind RNA which allows them to undergo this phase transition. So, here’s the classical example that was first… where it was first proposed that these membranous granules are indeed kind of a liquid composition. And this was carried out by Cliff Brangwynne when he was in Tony Hyman’s lab. And what you’re seeing here is the PGL-1 protein, labeled with… tagged with a GFP, a fluorescent protein. And you will see… this as an embryo… uhh… uhh… a one-cell embryo of C. elegans. And what you will see is that the granules in the front part of the embryo, in the anterior part of the embryo, will dissolve, and they will assemble into larger particles in the posterior half. And that will be the posterior cell. And that cell will actually give to the… give rise to the germ line in C. elegans. So, let’s play the movie. You can see how these granules are now associating. You can probably see here where there are some granules fusing and merging. And so that suggested that there was a liquid kind of property to them. They’re behaving like liquid droplets. And that means proteins are going in and out easily, and they can merge, sort of like putting oil into water. There are some ways of how we can assess the liquid nature, or the fluidity, of the molecules within the membrane compared to the outside. And so, two fluorescent methods are used for this: recovery after photobleaching, where we photobleach the granule and then ask how easily is the content of the surrounding now filling in and providing, then, florescent activity; and the other assay is called fluorescent loss in photobleaching, and here we are using photobleaching all the time in a particular spot, and if molecules from the outside are merging in, we would expect that, then, we’re losing fluorescence from the outside, because they are filling in and are getting bleached. And so, both of these aspects can actually be visualized when we’re going, now, into the Drosophila germ cells. And so, I want to just tell you something about the cytoplasmic and nuclear granules in Drosophila, before I show you the experiment using FRAP to show the mobility of the proteins in the granules. So, in the… during embryogenesis, we have these granules which are formed by Oskar and Vasa, and contain RNA. Then, as the germ cells form, we actually find particular granules in the nucleus, which are specific for the germ line. And then later, during oogenesis, we have another type of granule, and you can sort of see they’re surrounding each of the nurse cell nuclei, and those are called nuage. And so, here’s the experiment, where we are asking, if we’re bleaching germ granules in Drosophila, can they recover? And as a comparison, I’m showing you the results that were obtained for PGL-1. Indeed, we did these experiments because we wanted to repeat the experiments that Cliff Brangwynne did so we knew we were actually measuring the right thing. And so, as you can see, this would be how much fluorescence there was before the bleaching. And PGL-1 recovers about 80%. So, there’s a large mobile fraction which can move into the granule after… after the photobleaching. However, when we do this experiment with Oskar, there is a mobile fraction, but there’s also quite a large fraction — it’s about 60% — which is not mobile. So, we would say that probably the Oskar granules are not as liquidy as the PGL granules, and that there is perhaps more structure to them. And so, how can we get to the structure of the Oskar granules? And one experiment that got us to the structure of the granules was sort of almost a coincidence. We wanted to know something about the two forms of Oskar. There’s a long form and a short form. The short form does all the germ plasm assembly that I was telling you about. And the long form has been a little bit of a mystery to us. And so we thought, let’s just express it in tissue culture cells and see what it does. And what we found was that it was kind of everywhere in the cytoplasm. But then, as a control, we thought, let’s also express the short Oskar in the tissue culture cells, and there was the big surprise. Because short Oskar — in Schneider cells, which are Drosophila cells, or in HEK293 cells, which are mammalian cells — was on its own, without any germ plasm, without any other Drosophila proteins, able to form granules. And so this was really an exciting moment, because it really told us that we can now study the biophysical properties in a tissue culture system where we can add components or subtract components. What was important for the nuclear localization is actually the nuclear localization signal. And that sits in that domain, the wiggly domain, that I was describing earlier when I showed you the structure of Oskar, in that un… in that disordered structure. However, granule formation of short Oskar is not dependent on one particular region. We can delete each one of these regions and still form granules. Only when we delete the nuclear localization element will the granules now form in the cytoplasm. So, what can these granules do? Indeed, what was really nice was to show us also that they had, really, biological properties which resembled those properties it has in vivo… was that we asked VasaGFP… which normally is associated with Oskar, but when we express it in these tissue culture cells, it does not go into the nucleus. It actually barely forms granules. But when we expressed Vasa and Oskar, we got a colocalization, and Oskar dragged Vasa into the nucleus. So, we’re still seeing the same kinds of properties that we’re seeing in vivo. So, what are these granules? Indeed, Mahowald, many years… in the 70s, when he was doing an EM construction of the germ plasm and first told us about germ granules, he observed these nuclear granules, and he showed them. And the granules that we’re making in tissue culture cells have the same size. They’re actually quite big: they can be over a micrometer big. And… and they have… they look kind of hollow. And if we do a serial reconstruction, we realize that they are not a donut, which kind of… you may… but they are, really, a cream-filled donut. And so they are a hollow sphere. Although we don’t know… hollow is perhaps not the right word, because we don’t know what’s in the middle, but it’s electron less-dense. So, really, there is an organization to these granules. They are not just a concentrate which comes together, but they’re organized. And so, when we compare the germ granules in C. elegans to Drosophila, and in general think about the properties of granules, of these membraneless granules, there are a number of properties. First of all is that they’re pretty much round and membraneless, and that’s true for both. They have a self-driven assembly, and I just showed you the proof for that. The can demix from solution. I showed you that too, with the FRAP experiments. They’re much increased in concentration, and we’ll talk about that a little bit later also. They’re stable once in phase, and we know this with additional experiments that we’ve carried out for the nuclear granules in particular. They’re multivalent. I told you that there are multiple domains which can… which are required for the granule formation including the RNA binding domain of Oskar. And their stoichiometry is not defined. So, while I was telling you that Oskar dimers and Vasa dimers… form with Vasa to make a heterodimer, the other stoichiometry of the other proteins which are in the germ plasm are probably not as clearly defined. And that is another hallmark of germ granules. So, now finally, the last part of my presentation will be to really look at how the RN… RNAs are organized. And here, just like what I was just telling you, a major part is really that these granules are not just jelly. They… they… that demixed. They have a structure. And we will see this in particular when we’re looking at RNA localization, and that was quite a surprise. And so, there are RNAs in the… in the germ… in the germ granule. And here you see, for example, VasaGFP. This big one is one of those nuclear granules. You see how much smaller the cytoplasmic granules are. And then you see RNAs associated with these granules. And so, the germ plasm assembles during oogenesis, as I told you. Then we have that rotation during the dumping of the nurse cells. And during that time, these RNAs localize that are important for all these different functions. And Tatjana Trcek, a postdoc in my lab, really took it to herself to do a much more quantitative analysis, and look at where these germ granules RNAs are actually localized. And so, what you see from this image… and this image is actually not an image of proteins, but of RNAs. And you can pretty well see that there is sort of a line where you see a much more dense area of RNAs than in the rest of the embryo, which is the soma. And so, what Tatjana used for her quantitative analysis is single molecule fluorescent in situ hybridization. And in this case, it allows us to observe every molecule of RNA, and count every molecule of RNA. So, the RNA is labeled by little pieces of complementary RNA, and each piece is fluorescently labeled. And when we add, like, about… more than 40 of these little pieces going along the RNA, then we can see single molecules. And this dot here is a single molecule. When we get to the germ plasm, you can see it’s very crowded there. And so here, we are counting the molecules pretty much on the basis of how much fluorescence intensity there is, knowing what the single molecule intensity is. So, this was really exciting for me, because it’s the first time we were able to count the molecules. So, we know, now, how many molecules for nanos are in the… in the soma. And we get a concentration measurement of this too. But most important is to compare this future soma, where the RNA is localized… and indeed, I should say all of this RNA gets degraded as soon as the germ cells form. And only the RNA which is in the germ plasm, which gets incorporated into the germ cells, will persist. So, what is the concentration in the soma? And so, only 3% of the total RNA localizes to this region, which is about 0.1% of the volume of the embryo. And that leads to an incredible enrichment of this RNA. And that is the first point which also suggests that the RNAs are also assembling into clusters, and that these clusters are… similar to what we were observing with the proteins, are now having a much higher concentration, and can perhaps organize in different ways because they have a higher concentration. So, if you just think from a biochemical point of view, if you have molecules in solution, it’s gonna be very hard for those molecules to find each other. But when you concentrate them, very different biology can happen, very different interactions between molecules can happen. And that’s what I’m going to tell you about. When molecules are in a dense space, it is actually very hard to determine whether their colocalization is meaningful or whether it just happens because they are close. And so we had to develop a way of determining, in this dense space of the germ plasm, the colocalization of molecules. And what we did for this is we looked at the peak of fluorescence of two proteins, or two RNAs, or an RNA and a protein. And then we measured the distance of those two peaks. And our rationale was if that distance is pretty constant, then these two molecules are colocalized. However, if the distance varies a lot, then the two molecules, despite the fact that they are in the same region, they are not colocalized. And this measurement can be explained by the Pearson correlation coefficient a la Costes. And the… and I will show you that later. So, we can look at protein-protein. And when we look at Vasa and Oskar, here, and we just look at… with structural illumination microscopy, which allows us a resolution that is very good, close to super-resolution microscopy, we can see that pretty much even with just our eyes… here, without a formula… that pretty much all of the proteins are colocalized. However, when we looked at different RNAs, what we see is that some RNAs seem to be colocalized quite nicely, like cyclin B with Vasa, here. But then, other RNAs, like germ cell-less, are not really that well localized… colocalized with Vasa. And so here is the Pearson correlation coefficient that I was talking about. 1 is the maximal. That would be perfect correlation. And so we had to, of course, have a positive control for that. And the positive control would be how well would the same molecule come out, and of course the same molecule should be totally colocalized. And so what we did for this is we took the same RNA and labeled it with two different colored probes, and then compared those peaks. And this gives us the optimum, and of course it isn’t 1, because there’s always some variation. But that is basically telling us, this is as good as colocalization could be. We also had a negative control, and this is ccr4. We know ccr4 is a little bit enriched in the… in the germ plasm region, but it is clearly not localized. And you see there’s a lot of distributions, and it is not a constant… a constant distance from Vasa. So, everything else, now, we’re measuring with regard of Vasa as the… our guide for the center of the germ granule. And so, you are not surprised to see Oskar protein being close to 100% percent colocalization, as good as our positive control. They are partners, and so they are together. What was really surprising… and I want to take you back for a moment. So, for years, we had been doing in situ hybridization with more traditional probes. And all these RNAs looked identical. They were always enriched at the posterior pole, and that was… like we called them, germ plasm RNAs. And in situs were being done by EM, and they showed that they were actually localized in the granules. But… so, the big surprise was when we looked at how the RNAs related to the peaks of Vasa. What we realized was that they were organized. Some RNAs, like cyclin B and nanos RNA, were more in the center, while other RNAs, like germ cell-less or pgc, were more at the periphery. But they had a very precise orientation, and there was… they’re still colocalized. It’s not random, because they have a specific position. This suggested to us that the RNAs were organized with respect to the proteins. And it also told us that different RNAs formed clusters of higher intensity, which told us that there were multiple of these molecules, that these RNA clusters were organized. And the same RNAs were… obtained different positions. And to just illustrate this… so, the different RNAs can all be in one granule. And so we asked the question, if we just now draw this, just for the RNAs, we can actually triangulate and get an idea of how the RNAs relate to each other. Important for this localization is the 3′ UTR of the RNAs. So, that is how they get to the posterior pole. The 3′ UTR is also important to keep the RNA stable in the germ plasm, and lead to their degradation in the soma. And finally — and this is still a mystery for us — the 3′ UTR also determines when the RNAs are translated. So, nanos is translated in the germ plasm, but germ cell-less is translated as soon as the germ cells form. and cyclin B is actually not translated until the germ cells reach the gonad. So, what is it about the 3′ UTR that first of all gets them localized? And what is it that gets it organized and translated at particular times? I do have an answer… or I think I do have an answer… for the first question. I still don’t have an answer, really, for the second question. So, here’s a model of RNA localization. And the model of RNA localization can be described as a zip code model, where the RNA… in this case, the 3′ UTR, as we know most RNAs are localized through their 3′ UTR… there are a few exceptions… where the RNA has a particular tag, which would be a particular sequence or a particular structure, and that specific proteins bind to that tag, and they then bring the RNA to their location. So, the specific sequence, like illustrated here, has really not panned out very well. There are sequence-specific regions in the 3′ UTR that are bound by sequence-specific RNA binding proteins, but those turned out mostly to be involved in RNA translational regulation. The second model, that proteins… specific proteins bind to structures in the RNA, has been shown for a number of examples. And that’s often when RNA is packaged into particles, which are then transported in a microtubule-dependent way, for example Oskar getting to the posterior pole, or bicoid getting to the anterior pole. But our RNAs do not seem to behave this way. What we think, how we can describe those clusters best, is by the RNAs organizing with themselves. Because we have no evidence that there’s any specificity in the protein, because I showed you the RNAs were organized, but the proteins, Oskar and Vasa, they were everywhere, right? They didn’t have a specific localization. And so, what we would like to propose, and we have some evidence for this, is that the different RNAs self-associate. The proteins are required for them… to allow to be enriched, but then it is… the specificity comes really from the specific sequences of the RNAs. And that’s where the specificity problem can really be solved, because, remember, there are 200 RNAs. Are we gonna have 200 specific RNA binding proteins? I don’t know where they are, because we haven’t found them. And so, in this case, if the specificity comes from the RNA, we can really solve that problem. So, how could specificity come from the RNA? Now, think back at the high concentration? So, when RNAs are in high concentration, and when they have, then, a chance to engage in trans interaction rather than in cis interactions, as in kissing and pairing, where loops can connect in trans, or where stem-loops or sequences can open up and find each other in trans, where two molecules then can interact, that would be a mechanism of specificity which would be innate in the particular RNA. And so, what the model we are working with, at the moment, is that initially the proteins… because Oskar, Vasa, Aubergine, Tudor, they localize and start forming germ plasm before the RNAs even get there. So, they would be involved in seeding the RNA onto this protein scaffold. And then the RNAs would self-organize. What is nice about this is it explains, for example, how you can make a large granule if you have a lot of RNA, and that’s where we see a correlation. So, the more RNAs present, the larger the granule. And those granules are usually also in the center. This model, despite the fact it seems a little odd, because we are so protein-focused, often, always, is not the only existing one for the germ line of Drosophila. And there are some really neat recent examples, in other systems, where people have now discovered these RNA-RNA interactions as important. So, for example, in this filamentous fungus, Amy Gladfelter’s lab has shown that certain RNAs — in green and purple, here, and blue is the DNA — are self-associating, and she can actually show that this is due to the RNA structures that are recognizing each other. Similarly, Ron Vale’s lab has shown that RNAs which have these disease-associated repeats, where in the disease the repeats are longer… of certain nucleotides being repeated again and again… that they… when the repeats get very long, the RNAs can also self-associate. So, perhaps this is a more general model of how these specific RNAs can engage with the granules and where the granules, the protein part of the granule, is providing a scaffold for these interactions. So, I want to summarize this part of my presentation. I first want to tell you that Drosophila germ granules have physical properties similar to other membraneless granules. I told you that they are also more structured. And this of course tells us more about a granule which has many components, and which is acting in vivo. And obviously the granules will change. So, for example, the cytoplasmic granules we already know have other components than the nuclear granules, which are coexisting at the same time. Aubergine and Tudor are not in the nuclear granule. They’re only found in the cytoplasmic granule. Only the cytoplasmic granules seem to assemble these RNAs that I just showed you. We do not see the same RNAs with the nuclear granules. The granules are the site of protein and RNA localization. And the RNA localization is really what makes the functional aspect here. The germ granule effector RNAs encode all the proteins with the important functions for germ cells. And so the germ plasm is just allowing them to be concentrated and then translated there. And finally, the germ granules are organized in these homotypic clusters. And these RNA-RNA clusters are perhaps a new way of thinking about how specificity can be achieved for specific RNAs to organize within a granule. And it may also be… perhaps give us some idea of how they are controlled, either to be prevented from translation or activated for translation, because that’s of course, at the end, their functional aspect. And so with that, I want to thank, really, the major contributors. So, Tatjana Trcek has really contributed greatly to our analysis of RNA and the granule biophysics. And those experiments were also conducted, many of them, by Katie Kistler, with support from Thomas Hurd. I would also really like to mention our collaborators: Timothee Lionnet, who was at Janelia Farms and now joined us at NUY, and Hari Shroff, who allowed us to use his structured illumination micros… microscope, and they really contributed a lot in the analysis of these granules. And obviously Cores. And some of the drawings that I showed were made by Alexey Soshnev. And finally, there have been many people, throughout the years, who have worked in my lab on RNA. And obviously, all of this couldn’t be done without the funding agencies. And so I thank you very much for your attention. And I hope you join me for the second part, where I will be telling you more about how germ cells form, and some of the aspects of how they become different from the soma, and how they make sure that they don’t undergo the same deadly fate of the soma. Thank you very much for listening.