So in the previous lectures we've been thinking about specific epigenetic marks about the chromatin and how it's organised and whether it's densely packed or open. And how the chromatin is moved around and how particular epigenetic complexes are targeted. In this lecture what we'd like to thinkabout now is the three dimensional organisation of the nucleus. So, we have to condense down two meters of DNA into the nucleus. And so one would expect that all of the nuclear space would be taken up by chromosomes. But in fact that's not the case. We know that each of the chromosomes exists only within, or predominantly within its own particular territory, and I'll come to this in the next slide. And also that about half of the nuclear space is not actually taken up with chromosomes at all, but is rather made up of nuclear sub-compartments. There are many nuclear sub-compartments that we can think about: the nuclear lamina, the nuclear pore, the nucleolus, transcription factories, polycomb bodies, paraspeckles, splicing bodies and probably many more. In this lecture we're just going to think about these first four, the nuclear lamina, nuclear pore, nucleolus, and transcription factories, as well as thinking about those chromosome territories. So here on the left-hand side is an image of a nucleus of a male cell, a male fibroblast, where each chromosome is painted in a different color. So, you can make a chromosome paint by purifying a particular chromosome and then labelling it with a colour, and then through affinity it will bind to that particular chromosome when you put it back on to a nucleus. So, when you deconvolute this, this raw image here, you end up with a location of all of those chromosomes within that particular nucleus. And what's really striking is that each chromosome occupies its own territory. That is it's not spread out throughout the nucleus but rather it tends to be constrained in one region. What's also interesting is that these chromosome territories vary according to cell type and also developmental time. So they must have some functionality. It's not just to, you know, keep them restricted in space, so that they don't get tangled up, if you like. But it must have something more to do with gene expression. And what I think is really interesting is that, if you take and find out where translocations happen in cancer cells, so often in cancer you'll have crossing over of chromosomes in break points, and you'll have recombination occur where it shouldn't, these translocations. And these happen between neighbouring chromosomes in the cell of origin. In other words, what you tend to see is that, for example, chromosome 6 and 10 from a fibroblast might translocate with one another. But you're very unlikely to see translocation between chromosome 6, for example, and chromosome 3, because they're not physically located next to each other in space. And so I think this makes perfect sense, and really makes sense of these chromosome territories, that they must really exist, and this is how they, this is how they actually appear within the nucleus. Beyond the chromosome territories there are these nuclear sub-compartments. The four as I said that we'll talk about are transcription factories, and they are here in the centre of the nucleus shown in green. The nuclear lamina which is around the outside of the nucleus and the nuclear envelope, the nucleolus and the nuclear pore. So, while these chromosomes are within their particular territories within the nucleus, they can loop out to, small regions of these chromosomes can loop out and visit one of these nuclear sub-compartments. It can either move to a transcription factory to be transcribed as the name suggests, to the nucleolus, to the nuclear pore, or indeed to the nuclear lamina. So, we'll go through each of these in turn to describe their function. The first of these nuclear compartments we'll think about is the nuclear periphery. So, in general, association around the exterior of the nucleus, the nuclear periphery, is associated with silencing and in particular with facultative heterochromatin, that is the genes that can differ in their silencing between cell types. One example of this is actually the inactive X chromosome, as usual. But in the interior we find active genes and so for example in the transcription factories, so the centre of the nucleus is associated with euchromatin, and the exterior or the periphery of the nucleus is associated with heterochromatin. On the inside of the nuclear envelope, the inner nuclear membrane, we have associated here nuclear lamins. And these nuclear lamins are proteins that other chromosomes can attach to. What we know about this is that the regions that are attached to the lamina tend to be lowly expressed, as I said. But we also know that if you force attachments to the lamina, this results in silencing. And this is really suggestive of the attachment to the nuclear membrane causing silencing rather than being a consequence. We know that the genes that are attached to the nuclear lamina stay relatively consistent between cell types, so 70 to 90% are the same but that this can vary upon something like activation of the cell. So some cells, like immune cells, can be activated and be responding to infection for example. And in this case you have a very rapid and dramatic change in the gene expression profile. And this will also result in a change in which genes are associated with the nuclear periphery. What I think is really interesting is that if you reposition a gene away from the nuclear periphery, this of course is associated with transcriptional activity. But the relocation away from the nuclear periphery and moving in towards the centre needs to happen before you can actually have transcription ensue. So, so why is this? So why do we have the nuclear lamina being associated with heterochromatin, with silencing of transcription? Could it be that at the nuclear periphery you have a concentration of silencing factors? This is something we don't know very much about yet. Or could it just be that it's out of the way of the transcription factories, out of the way of these regions that are actually where transcription is occurring, based on their name. These are what we'll discuss next. So, what are these transcription factories which I've just mentioned? They are bodies, there are multiple ones within each nucleus, and I'm just showing one here. And within these bodies, they have a very high concentration of RNA polymerase. In mammals, there are three types of RNA polymerase, RNA polymerase one, two and three. And it's RNA polymerase two, which is used to transcribe just your regular genes. So, in these transcription factories, we have more than a thousand-fold higher RNA Pol II concentration than you do anywhere else in the nucleus. And what's interesting is that in these transcription factories, the RNA polymerase two itself appears to be static, but it's the DNA, which is fed through the RNA Pol II to be transcribed. We also know, in addition to having RNA Pol II in the transcription factories, they also have a dense concentration of transcription factors. There, we know there are a large number of transcription factors that are encoded, and different transcription factories have variant concentrations of particular transcription factors. We know that correlated genes, that is those that are regulated together or need to be expressed at the same time, share the same factory. And this of course is a very efficient process because it means that we don't need to have such high concentration of transcription factors in the nucleus. Rather it's just a high local concentration because they're all found together in the transcription factory. And this makes sense of why we know transcription factors in fact are really very, very lowly expressed genes. They don't need to be highly expressed when you look at the whole nucleus, so long as there's enough of them at one particular site. One of the questions is then, well, are these transcriptions factory, factories stable? Or do they rapidly self associate when required? And this is one of those questions that we still don't understand. If we move on now to the nuclear pore. So the nuclear pore is exactly as it sounds. It's the pore to allow things to get out of the nucleus or to come into the nucleus. But some genes are in fact tethered to the pore, appear to be tethered to the pore or located very close to the pore, and these tend to be active genes. It's euchromatin, actively transcribed chromatin. So it's not just found in mammals in this case but, in fact, it's found in yeast, in flies, in basically all organisms between yeast and mammals, so there's must be some reason that there can be euchromatic genes associated near the nuclear pore. And while we don't really know yet what this reason is, we do know that, or we can suggest that maybe it's because it just allows fast export. So, if something is being transcribed at the nuclear pore, it can rapidly get out and get to the ribosomes, where it will be translated. The last of these nuclear sub-compartments is the nucleolus. So this nucleolus is a relatively large large nuclear body. And we can see it, and the reason that it can even be seen sometimes with staining with DAPI, so staining with this DNA dye is because it has all of the ribosomal DNA repeats, or the rDNA. So the ribosomal DNA is transcribed into the ribosomal RNA and this is used in the ribosomes that are important for translating protein or the only thing really important for translating protein. So it's the rDNA repeats that are found on several different chromosomes that are all clustered together in this nucleolus. And the reason they cluster there is because this is where they are transcribed. So it's kind of like the transcription factory, but a transcription factory specific for the ribosomal DNA made, to be made into the ribosomal RNA. In this case it's a different RNA polymerase; it's RNA polymerase one. And RNA polymerase one just works in the ribosomal DNA. What we know is that the transcription in the nucleosome, sorry in the nucleolus is really important in this case and so here you have the transcription and the processing and even subunit assembly that's all happening in this nucleosome. So it's really important for the function of the ribosome in the end. In this case, in addition to the actual nucleolus, which, if we consider this the nucleolus, around the edge of the nucleolus is what's considered the peri-nucleolar space. And it's in this peri-nucleolar space that we can really consider this almost as a separate nuclear, sub-nuclear compartment. In this region there's also, rather than having RNA, lots of RNA Pol I which is found inside the nucleolus. But rather around the outside what we have is a concentration of RNA polymerase three. And RNA polymerase three is important for transcribing the transfer RNA genes or the tRNA genes. So, again, we don't really know why this happens in the peri-nucleolar space, but potentially, there's some relevance to tRNAs which are required, of course, for translation, being located nearby to the rRNAs, which are making up the ribosome. Both of which are involved in translation. So now what I'd like to do is just come to a very brief summary of what we've done in these first two weeks of this course. First of all I went through the various epigenetic marks that are found with active genes or euchromatin and those that are found with inactive genes that are found in heterochromatin. So, you'll remember that in active locus we find a particular set of histone marks, so H3K4 methylation and here, shown by the diamond, the acetylation of histones and we find no methylation of the DNA. We know that these nucleosomes are relatively sparsely located and that's because of the function of chromatin remodelling complexes which are opening up the chromosomes, chromatin. Whereas, if we're looking at an inactive locus, where it's not being transcribed, the nucleosomes are more densely packed again, through the function of the chromatin remodelling complexes. And you can, we know about a set of histone marks and DNA methylation that are found in this case. So H3K9 methylation, H3K27 methylation, and of course, DNA methylation. We spoke briefly about the fact that these different epigenetic marks can act as docking sights for other chromatin proteins. And these chromatin proteins can bring about wider changes in the in more epigenetic marks or they can remodel the nucleosomes that are found there, so they may actually the histone marks may bring in chromatin remodelling factors. We went briefly through histone variants and how each of their specific features allowed their particular functions. For example, DNA repair with this gamma H2AX stain shown here. And then we thought about the noncoding RNAs. The piRNAs which could direct DNA methylation at transposable elements, and the long non-coding RNAs, which we’re able to direct epigenetic modifiers to particular sites in the genome. And finally what we've gone through now is the nuclear architecture, and the subnuclear compartment, which can contribute to gene silencing, or gene activation. So by taking into account all of this knowledge that we've acquired over the last two weeks, what we want to do in the subsequent four weeks is go through particular examples of epigenetic control, and use what you now understand about epigenetic control to describe what happens in each of those instances. So now let's again consider this heritability of epigenetic marks. We know that by definition an epigenetic mark needs to be mitotically heritable. But we don't necessarily know all, everything about whether all of the marks that I've spoken about over the last two weeks in fact are mitotically heritable. What I mentioned last week at the end of week one, was that DNA methylation is mitotically heritable and we know this happens through the action of DNMT1 and because the de-methylases are not expressed very often. Histone methylation we know a little bit about how it can be mitotically heritable predominately through the study of polycomb repressive complex two, which lays down histone methylation at histone 3 lysine 27. We know this because if you consider strands of DNA when they're being replicated and the nucleosomes which are found on the DNA, we know that PRC2 can sit at the replication fork itself. So when it's sitting at the replication fork itself, it can both recognise the parental nucleosome and recognise that it has methylation H3K27. But it can also lay it down onto newly forming histones. We don't know many of the details for how this occurs, but we do know that because PRC2 both recognises the methylated residues, and unmethylated residues, this is partly how it works. And because of its location to the replication forks. Histone acetylation is a bit different. Histone acetylation we know is in fact not mitotically heritable. That's partly because the histone de-acetylases, and the histone acetyl-transferases, are both present most of the time. And in fact histone acetylation should be rather considered as a chromatin mark, instead of an epigenetic mark. We also know it rapidly changes, even throughout one cell cycle, and it's not maintained. And the best example of this is that histone acetylation changes during the diurnal cycle, so, in other words, with circadian rhythm. So the cells may not themselves be dividing, but because you're going through, a night cycle and a day cycle, histone acetylation will change during these periods. We know histone variants can be mitotically heritable because they're often loaded after replication. And for long non coding RNAs we also know they can be mitotically heritable. And I think the best and probably most exciting example of this is known for Xist, that long non RNA coding involved in X inactivation. Here we know if you take chromosomes that are found in the dividing cell at meta-phase you can still find Xist coating the entire inactive X chromosome and so it's passed from parent to daughter because it's still bound on that chromosome itself. piRNAs are a little bit different. We know that piRNAs the effect is, the piRNAs, the effect of them is to set up a cascade of epigenetic events. And so the epigenetic marks that they establish are themselves mitotically heritable. So, for example, piRNAs are able to direct DNA methylation and we know DNA methylation is heritable. And we just spoke about nuclear architecture. We don't yet know how nuclear architecture can be mitotically heritable but it seems that it likely will be because nuclear architecture within one particular cell type tends to be maintained. So next week we are going to start to think about all of these epigenetic marks that we've just learnt about and how they bring about X chromosome inactivation.
