Okay, so we're now going to consider a little bit more about DNA methylation. So where does it occur? We've mentioned in the last lecture that it can occur at CpG islands, although in general, they tend to be unmethylated. But the other regions of the genome where you find DNA methylation, in fact, where they tend to be methylated, are intergenic intervals and their repetitive elements. So here in this picture we've got the CpG island upstream of the gene, but then the intergenic interval, here, we see have these little lollipops. And when they're black and colored in that tends to indicate that they're methylated, whereas when they're up here and white and open that indicates unmethylated. And this is a very kind of a universal way of showing CpGs in this field. So you can see that these repetitive elements and the intergenic regions tend to be black and methylated in this picture. So, if we think about where the CpGs are, it's important to remember that actually there are fewer of them than would you expect by chance in the genome. And that's because DNA methylation is actually mutagenic. So here we've got a picture of the Cytosine, a 5-methyl Cytosine, and Thymine. And what we know is if we take this 5-methyl Cytosine, this amine group at the top of the ring can relatively easily be lost, and this will make the 5-methyl Cytosine convert into a Thymine. You can find more TG residues or dinucleotides in the genome than you might expect and fewer CG residues than you'd expect by chance because of this mutagenic event. So what's the function of DNA methylation at intergenic regions? Well, we think that it's there to maintain genomic integrity. It's clearly not usually silencing the expression of a normal gene, but we do know if you look at cells that don't contain DNA methyltransferase one, this maintenance methyltransferase, then the cells exhibit genomic instability. What I mean by that is if we consider the karyotype of a normal cell, here just a very simple version with four chromosomes of different colors, then when there's genomic instability, you can have many things occurring. For example deletions, where we've lost a bit of this green chromosome. Insertions, where we've gained a bit of red chromosome. Even duplications of whole chromosomes, here shown as an extra red chromosome. And as well as translocations between chromosomes, where the black and the blue chromosome have swapped over some pieces. So without having the appropriate methylation at these intergenic intervals, this can be what results. We also know that methylation does in this case also seem to do some silencing of expression. But it's not a traditional promoter that's silenced but rather a cryptic promoter or a cryptic splice site. So if you consider that you have a DNA, sequence and there might be a gene up here which has some exons, and there might be a gene down here which has its own set of exons. In the middle, so these make both be expressed in the same direction, and so they're the promoters shown here with he arrows. In the middle, you can have a small and possibly not particularly powerful promoter heading in the other direction. It might be there by chance, or it might be there associated with a different gene. But what we don't want to happen is for this to have RNA preliminaries, too, loaded on. And also this to have RNA preliminaries too layered on. Because then what happens is they both passage across and they bump into each other. These two RNA preliminaries, two molecules, bump into each other and they result in transcriptional interference. This is particularly a problem if it happened within the gene say, for example, here. Because then you won't have the hull of the first transcript made at all. We also know that you can have particular splice sites that can be silenced by this intergenic methylation. So splice sites have a particular DNA sequence that they are recognized by the splicing machinery. But if this DNA sequence is found at other regions within a gene, then, that you perhaps don't want to use, then again that will need to be silenced by DNA methylation as well. If we now think about what the function is of DNA methylation at repetitive elements, we first need to think about what is a repetitive element. So, in this case, the DNA methylation also seems to be there to maintain genomic integrity, or to maintain genomic stability. So drawn on the bottom of this slide is one example of a repetitive element. And this is an intracisternal A particle, or an IAP. In this case, you can see these orange, long terminal repeats at each end, and inside are the proteins that are required for the function of the repetitive element. In general, what repetitive elements do is they make a copy of themselves, they jump out and paste themselves into somewhere else in the genome. So this can either be a copy and paste mechanism, or a cut and paste mechanism, meaning they do or don't leave behind a copy of themselves where they were. Clearly this is mutagenic because if you're going to paste yourself somewhere else in the genome, who knows what other genes you may disrupt when that happens. To be able to allow this mechanism of copying and pasting somewhere else, you need particular genes that retro transpose and or transposition needs to have particular functions to be able to allow this to happen. And this are contained in the gag and pol protein strain here. So these they need to express, they need to transcribe and then translate. So these long terminal repeat contains a very, very strong promoter. This strong promoter is associated with some CpGs that are shown here in the usual lollipop structure, just here. But because the long terminal repeat at the other end also contains this very strong promoter, and similarly contains these CpG dinucleotides, here, if this promoter were active, it will actually read out into the down stream genes. So we know this can be mutagenic because you have transcription of the neighboring genes when you're not suppose to. This long term repeat has such a strong promoter that it will on in every cell at all times. So clearly, there's a drive in the cell to silence these repetitive elements. Firstly, to be able to prevent that mutagenic transposition that can occur. So if we can methylate these CpGs here then that transcription should be silenced. They won't make the gag and pol proteins, and they won't make the jump around the genome. Secondly, because DNA methylation is mutagenic, we know that if it's there then you're more likely to have the methylcytosines mutated into thymines. And this may also over evolutionary time prevent transposition because the mutations will mean perhaps the promoters don't work anymore. But, similar to what I mentioned for the intergenic regions, silencing of these repeats will also avoid transcriptional interference. This is partly what I've shown here where the LTR will lead to transcription downstream from the strong promoters. And even what's shown just here is a cryptic promoter so the LTRs even have promoter activity leading out the other direction. So, as you can imagine, many of the surrounding regions can be disrupted for many different reasons. There can be a genomic disruption because of the insertion of a large piece of DNA or even a transcriptional disruption because of these strong promoters. Finally, we think that methylation of these repeats can also prevent illegitimate recombination. Because there are a lot of these repeats spread throughout the genome, if they align, you would normally allow, recombination would occur between similar repeats. However, we know for a combination to occur we need to have an open creamish instructor. So this is something that's been learned only in the last few years in the field, but through methylation, this will allow compaction of this particular chromatin and recombination is much less likely to occur. So based on this silencing of the repeats, the genome defense model was proposed in the 1990s by Professor Tim Bestor. So this is based on the premise that if DNA methylation is mutagenic, then really there must also be a benefit for it to still occur in the mammalian system. And so he proposes that the primary function of DNA methylation is to protect the genome from these transposable elements, and this is certainly predominantly where DNA methylation is found. Because DNA methylation very rarely occurs at CpG islands, it mostly occurs at repeats or perhaps at intergenic intervals. So it's interesting to mention at this point what happens with DNA methylation in cancer. So, in cancer, rather than having the same sort of DNA methylation makeup as you do in a normal cell, they're actually global changes in DNA methylation. So these were discovered back in the 1980s, and in fact, they're a hallmark of cancer. They've been found for every cancer that's ever been studied. They have aberrant DNA methylation. And what happens in the case of cancer is that these intergenic intervals and their repetitive elements now tend to be unmethylated, and this means that as a whole in the genome, you tend to see hypomethylation. We also find that in some particular instances that CpG islands are hypermethylated, that is they're methylated when they shouldn't be, and we'll deal with each of these further in the lectures on epigenetics and cancer. But what this results in, in cancer, is we know that in cancer you tend to see genomic instability. You tend to see these illegitimate deletions, insertions, or reciprocal translocations and even duplications of chromosomes, or loss of chromosomes. So we think that this DNA methylation is hypomethylation at the intergenic intervals and their repetitive elements actually leads to, or helps to bring about that genomic instability in the instances of cancer. So while we've spoken about how DNA methylation can be laid down by the DNA methyltrasferases, we also know that DNA methylation needs to be removed at some particular times in development. So originally it was thought that this could only happen because you wouldn't have DNMT1 around. So, if DNA methyltransferase one, this DNA methyltransferase which recognizes the hemi-methylated DNA upon replication, if it wasn't in the nucleus, or it wasn't being expressed, it wasn't being made by the cell, then this would passively lead to demethylation. So this happens, we know that this DNA demethylation happens in early development and in primordial germ cell development, and at later times in differentiation. And in each of these cases, it's really essential for this demethylation to occur for cellular function. But it's been recently shown about ten or more years ago that this could actually happen without having DNA replication. So this passive demethylation, which could occur without DNMT1, requires that the DNA be replicated, and then you fail to maintain the methylation. And so eventually, with replication, you passively dilute out the DNA methylation. But if this can occur without DNA replication, it must, by definition, be an active process. So, as I said, passive demethylation I just described, can happen through this dilution effect. But active demethylation has now also been discovered. And it's not a simple removal of the methyl group, just a chopping off with scissors if you like of the methyl group, because this carbon-carbon bond is actually very difficult to remove. Instead, it's an enzymatic removal of the methyl group via several chemical intermediates. So this can happen in several different ways. But it seems to involve predominantly the TET proteins and also AID. We're not going to go into the details of this, but we now understand a little bit more about how active DNA demethylation can occur. So in summary then of DNA methylation, we know that it occurs predominantly at CpG dinucleotides in mammals. It's associated with gene silencing if it's found at the promoters. It helps to maintain genomic stability in several different ways. It's laid down by the DNA methyltransferases, and it's mitotically heritable because of the features of one of these, DNMT1. And it can be removed either passively or actively through the TET proteins We also know, which I haven't mentioned so far, that DNA methylation is essential for life, essential for viability of a cell. And we know this because DNA methyltransferase knockouts, that is mice that are made without, where we've removed the methyltransferase one gene, for example, die in utero just around mid gestation. And so they're essential, the silencing function that DNA methylation performs is essential to be able to make a whole embryo. So now that we've thought about how DNA methylation, that where DNA methylation is found in the genome and what its function might be, I'd just like to come back to this idea of the heritability of DNA methylation. So you remember that what I've said is the definition of epigenetics is that you have mitotically heritable changes in gene expression that don't alter DNA sequence. So really for DNA methylation to be considered as a bona fide epigenetic mark it needs to be mitotically heritable. So how is this mitotic heritability achieved? Let's just remember what we've learnt over the course of this week. This is brought about because of two reasons. One, the first is that we have DNMT1, the enzyme DNA methyltransferase one, which recognizes hemi-methylated DNA. So here in the picture, what I'm showing you is that you have your normal DNA being replicated and then when it's replicated you have these two daughter strands, these new strands of DNA that are shown in green. And they're pairing with each parental strand. And so this is the hemi-methylated DNA. So DNMT1, here shown in red, then recognizes this hemi-methylated DNA and restores DNA methylation to each of these new daughter strands, so that you have now two copies of the DNA, and they're methylated on both strands just like the original was. So this is how we have this mitotic memory of DNA methylation. But what I also just told you is that in addition to DNMT1 we have TET proteins and these TET proteins are actually active demethylases so they remove the methyl groups. So how then can you have mitotic heritability of DNA methylation if the TET proteins are around cutting off these methyl groups? Through a multi-step process. Well, the reason this works is that not only do you have DNMT1 around almost all the time, but the TET proteins are in fact, very rarely expressed. They're very rarely found being made into their protein products. Instead, TET proteins are really only being expressed in primordial germ cells and again in embryonic stem cells. And so the predominate mechanism is that of DNMT1, and the TET proteins really are found very infrequently. So next week we'll come to consider other epigenetic marks including histone modifications and other events like long noncoding RNAs and piRNAs, and how these contribute to epigenetic control.
