Hello, I'm Dave Bartel, I'm a member of the Whitehead Institute, a professor of biology at MIT and an investigator of the Howard Hughes Medical Institute. And today I want to tell you about these microRNAs. MicroRNAs are these small RNAs that are processed from these longer transcripts that can fold back on themselves to make these hairpin structures. And these hairpins are formed by Watson-Crick pairing, A's and U's, G's and C's, sometimes G's and U's, between the two arms of the hairpin. And once processed from this longer molecule, these small RNAs are much too short to code for proteins, but they have other very important roles in the cell, and that is to regulate genes. Now, we know that gene regulation is very important in part because this is what makes cells different from each other. Our bodies have many different cell types, but all these cells have the same genes, they all came from the same zygote. So, what makes a nerve cell different from a muscle cell different from a fat cell? Well, it's not the genes, they have the same genes, but it's how these genes are used. Where some genes are turned off in one cell but not in another, some are on, and you have different amounts of gene expression in these different cells. And that's in large part what makes them different from each other. And so we now know that microRNAs are regulating most of the human genes. And, of course, this regulation is important for not only what makes a normal cell, but also a disease cell, in a cancer cell. And so like many discoveries that have important implications for human biology and disease, the very first microRNA was not actually discovered in humans, but instead it was discovered in a model organism, in this case it was the model organism for nematodes, the C. elegans, which is often used to study animal development. And back in 1993, Victor Ambros’ lab reported this small RNA called the lin-4 RNA. And that RNA, they concluded, with help from Gary Ruvkun's lab, that the lin-4 RNA regulates the lin-14 messenger RNA. And that causes less protein to be made from the lin-14 messenger RNA, and they concluded that this happens through an interaction between the lin-4 RNA and the lin-14 messenger RNA. Seven years later, Gary Ruvkun's lab found a second small RNA also involved in the timing of development. Here, going for lin-4/lin-14, going from larval stage 1 to 2. And here going from the cell divisions of larval stage 4 to the adult. So, then, Ruvkun's lab found that this let-7 RNA, which is important going from larval stage 4 to the adult, is also found in flies, and in humans, and other animals. And it's also expressed kind of at later stages and so the idea was that it may be also regulating the timing of development here and there might be other small RNAs regulating the timing of development. And so then, a year later, Tom Tuschl's lab, my lab and Victor Ambros' lab found that these small RNAs were actually part of a much larger set of RNAs and that these RNAs that were found were sometimes temporally expressed in development but most of the time, not, and so presumably they were just in specific tissues or cell types and so probably playing roles apart from the timing of development. So, at this point, because we'd found these things molecularly, we really didn't know which genes they were regulating, didn't know what processes they might be involved in, but what we did know is that they were small. Ok, and so we decided to call them microRNAs. So, this is when people got very excited about this class of small RNAs because it was clear that there were going to be hundreds of different microRNAs in humans, over a hundred in worms and in flies, and these microRNAs weren't just regulating genes important in the timing of development, but they could be involved in regulating any gene, really, any gene that a biologist might be interested in could actually be the target of a microRNA. And so many labs started working on these small RNAs and there's actually now been over 25,000 papers that have been published on microRNAs. And here I just have a word cloud of words taken from the abstracts of these papers. Many, many important discoveries. And so in the rest of this talk I just want to give you a brief outline of how these microRNAs are made, how they recognize their targets and what they're doing. So, how are they made? Well, in animals, the key three proteins are the Argonaute protein, which is the, ultimately where the microRNA is going to end up, which we abbreviate this AGO, and there are also these two endonucleases: Drosha and Dicer. And so the first thing that happens is that the microRNA is transcribed as part of a much larger primary transcript, and it's transcribed by the polII polymerase, the same polymerase that makes messenger RNAs. And then while still in the nucleus, it's recognized by Drosha and its partner, Pasha. And together Drosha and Pasha recognize the very, this part of the hairpin, and they cleave, the Drosha cleaves, right about one helical turn from the base of the hairpin. Ok, and then, that releases the pre-microRNA hairpin, which then exits the nucleus, with the help of the exportin 5 complex, and then there, in the cytoplasm, it encounters Dicer, which lops off the loop, and now you have a pre-microRNA duplex. And then one of the strands of this duplex is loaded into the Argonaute protein, abbreviated AGO here, to make the silencing complex. Now, exactly which strand goes into the silencing complex, and how that happens, is a mystery, but one important clue is that the strand that's most likely to go into the duplex is the one that's least paired here at the 5' end, it has the least stable pairing at its 5' end. Now, in some cases, they'll have an equal propensity to go into the silencing complex, and both of them can target genes, but most of the time, there's a very strong propensity for one of these strands to make it into the silencing complex, and from there, the microRNA directs this Argonaute to target messenger RNAs, and other RNAs. Now, these primary transcripts actually come in three different flavors; there's some cases where you just have one hairpin made from the gene. In other cases, you have three. And we consider these actually three different microRNA genes, even though they come from the same transcript. So, it's sort of like a polycistronic transcript. And then in other cases, the primary transcript of the microRNA is also a pre-messenger RNA where it is spliced to give you the mature cDNA, I'm sorry, the mature mRNA, which is then translated into protein, but then, the intron contains this hairpin, which is recognized by Drosha and goes into the microRNA pathway. So, you can have these multi-functional transcripts. Now, there are some cases where the processing here actually skips either the Drosha or the Dicer step. There's an important microRNA involved in red blood cell development that doesn't use Dicer for its processing, actually, the hairpin goes immediately into Argonaute and that Argonaute can help process it into the mature RNA. There's other classes that avoid Drosha processing, it manages to bypass the Drosha processing and one type of these are cases where you actually have the primary transcript, the primary transcript, again, the microRNA is within an intron, but in this case, the splicing machinery actually defines the ends of the pre-microRNA hairpin, so that after this intron is spliced out and you have this lariat, when that lariat is debranched, it can immediately fold into a pre-microRNA hairpin, without any further processing and then these then are recognized by Dicer and go into the pathway. So, regardless of how these pre-microRNA hairpins, the important part is that the microRNA pathway goes through these short hairpins. So, and that's different than other RNA silencing pathways. For instance, we have the RNA pathway, where you also have dicer, and there's also Argonaute, ok, but in this case, the substrate that Dicer uses is very different. Instead of being this short hairpin, it's this long duplex. Ok, and a duplex can be made from two strands coming together, and pairing very extensively for a long distance. Or it can be made from a long transcript folding back on itself to make a hairpin. Either way, you get many of these small duplexes, which are called small-interfering RNA duplexes from a single precursor transcript. And then because of the way these long double-stranded RNAs are made, these guide RNAs, the siRNAs in the RNAi pathway, often go back and silence the same types of loci, not always, but often, go back to silence the same types of loci from which they came, so this is a very effective way that cells have to silence viruses, transgenes, transposons, et cetera. But it's different than the microRNA pathway, and the important difference is not so much what these RNAs do once they're in the silencing complex, but where they come from, in this case, long double-stranded RNA and in this case this shorter pre-microRNA hairpin. Well, many of these microRNAs, then, are conserved in different species. So, an example of that is miR-1. MiR-1 is found in human muscle, so it's found in your muscle, in your heart, in brown fat, and it's also found in the muscle of flies and worms. And the, although the hairpins, here, differ quite a bit, what's the same in each of these cases is the base pairing as well as the mature microRNA that's being made. You can see that this is, these RNAs coming from human, fly and worms, are actually very similar throughout their sequence, actually only differing a little bit here at the 3' end. So, presumably, this microRNA, miR-1, was in the last common ancestor of humans, flies and worms and presumably it's been playing important roles in the muscle ever since then. You can also find related microRNAs within a species. So, again, taking the miR-1 as an example, humans have three different members of the miR-1 family. Two of them make the identical microRNA and then there's a third member which is very similar here at the very 5' end and differs here in the middle and 3' end. And we group these microRNAs into families really based on identity of nucleotides 2 through 8, this is what really is crucial, and you'll understand why when I talk about the targeting. So, we have three members of the miR-1 family, flies have one, worms have two. And these families, and these microRNAs, again, are very often conserved in different species. So, as humans, we have at least 277 genes that are making microRNAs that are conserved in other mammals. And these, like the miR-1, fall into families, and so these 277 genes fall into 153 conserved families. Now, we also have many, hundreds, of non-conserved microRNA genes. These, some of them, might be playing important species-specific functions in humans, especially those that are expressed at higher levels. But most of the non-conserved microRNAs are actually expressed at very low levels, and that's why I can't say exactly how many there are, because some of them haven't been found, they're just expressed at too low of levels to have been detected. But, and so it's kind of unclear, those that are expressed at very low levels and whose sequence doesn't seem to be important in evolution, what they're doing, you know. One idea is that they're very recently emergent microRNAs, and haven't yet found a target that would somehow impart some sort of fitness advantage if it was regulated. And it's possible that, and probable actually, that many of them will actually disappear before they find a biologically relevant target. So, some of these probably occurring very transiently in evolution, without functions, but again, a few of them probably playing important species-specific functions, particularly those at higher expression levels. Well, of the 53 conserved families that we have in the mammals, 87 of them are also found in zebrafish. Flies, the model fly, Drosophila melanogaster, has 119 conserved microRNA genes and these fall into 94 conserved families. Worms have 108 conserved microRNA genes, falling into 59 families, and of those that are conserved here to zebrafish, from human to zebrafish, 33 of them are actually found also in the fly or the worm, or both. Ok, and that, again, that's like miR-1, that means that these have probably been playing important roles really since the beginning of bilaterian animals. And there's even one that's found in a radial, symmetric animal, the sea anemone, and that one, of the 40 microRNA genes that we found in the sea anemone, one of them is conserved to humans and flies, and worms. Sponge also has microRNAs, we found 8 in sponge, but these are, none of these really are related to those in these other animals, so it's unclear whether microRNAs emerged independently in the sponge, or if they were there in the last common ancestor. We actually think they were there in the last common ancestor because sponge also has the unique proteins involved in microRNA processing, the Drosha and the Pasha enzyme and so probably in the last common ancestor, but probably not too far before then, because more deeply branching lineages don't seem to have microRNAs. Well, what about in plants? Very similar story. The model plant often used to study dicots is Arabidopsis and it has 90 conserved microRNAs and these fall into 20 conserved families. Conserved both from monocots, rice, to dicots, Arabidopsis. And of these 20 conserved families, 12 of them are found even in these deeply branching land plants, such as the moss. Green algae also has microRNAs, at least 20 genes have been found there, this would be in Chlamydomonas, and, but these Chlamydamonas microRNAs are not related to the land plant microRNA. So, again, it's unclear whether microRNAs emerged independently in this green alga, and in the land plants, or whether the last common ancestor had microRNAs, but because there's so many differences between the processing of microRNAs in plants compared to the processing, and how microRNAs are made, in animals it's thought actually, and I think very clear, that microRNAs emerged independently in the plants and in the animals. And they are even microRNA-like molecules in certain fungi, like Neurospora. So, again, probably emerging independently there. And in each of these cases, emerging from the RNAi pathway, the RNAi pathway, this pathway that's very important for, in many different lineages, spread throughout most of eukaryotic evolution. This RNAi pathway involved in transposon silencing, virus silencing, et cetera, has probably been around since the last common ancestor of the eukaryotes, and in multiple cases, of plants, animals and in other cases, giving rise to this microRNA family where you have these small RNAs made from these small hairpin precursors. So, what then, are all of these little RNAs doing? And, we know that they're conserved, evolution must be preserving them for some reason, but what might they be doing? And so in order to try to sort this out, our approach has been to try to find reliable methods to predict the targets of the microRNAs. What are the genes that they're regulating? And this was pretty straightforward in plants because for the plants, we found that microRNAs often, conserved microRNAs, have very extensive and conserved pairing to a few plant messenger RNAs. And this type of conserved pairing, based on what was known already from silencing pathways in other instances, is predicted to give slicing of the messenger RNA, it's predicted that the Argonaute protein is going to slice the messenger RNA in half. And this killing of the messenger, of course, is a very effective way of reducing the amount of protein from this messenger RNA. And my favorite example of this is a case of the PHABULOSA gene. So, this is a gene that was known to be important in plant development because mutations in this gene caused the leaves to have this radial symmetry rather than having the top and the bottom, the plant was very stunted with these radial, symmetric leaves. So, PHABULOSA and its relatives had these mutations and it turns out that these mutations, like this G to A mutation, which originally were thought to be important because they changed the protein function, they actually happen to fall within this region that was pairing to miR166. And so when we saw this pairing to miR166 overlapping with these mutations, we thought, well, maybe the reason for this phenotype isn't because the protein is changed, because this A, as you can see here, changes this GGU, glycine codon, to a GAU, aspartate codon, but maybe that's not the reason that these plants are not developing properly. Maybe the reason is because this mutation disrupts the pairing to the microRNA. And so to test that, what we did is to change, instead of changing this G to an A, we changed the U to an A, because that still codes for glycine, and when we did that, we saw that now, even though the protein is the same, the plant still has this developmental phenotype where, and presumably, that's now because the microRNA cannot pair as well to its target. So, this showed that the regulation of this PHABULOSA mRNA by this microRNA, miR166, is very important in plant development. And these types of experiments had been done, actually previously, with a different family member of this family and with other targets of the plant microRNAs, and now many, many instances of microRNAs playing important roles in plant development have been shown. And part of that is because so many of the targets of the plant microRNAs have known or suspected roles in development. When we look at the 20 conserved microRNA families in plants, they have conserved pairing to 90 unique target genes in Arabidopsis. And 72 of those have known or suspected roles in development, including 65 transcription factors with suspected roles, or known roles, in development, like this PHABULOSA gene. And so, a strong enrichment here for targeting mRNAs with roles in development, and microRNAs playing important roles in that process, but then there are other types of targets that are also found to be conserved in plants. Well, that's sort of the picture that we have in plants. What about animals? Well, in animals what we see is that there are a couple dozen cases where we have very extensive pairing between the microRNA and the mRNA and that leads to slicing of the mRNA, just like what's seen in plants. And in fact, that's the reason that these siRNA duplexes that experimentalists will sometimes synthesize and deliver to human cells and other animal cells work. It's because they're recognized as these pre-microRNA duplexes and enter this pathway and give you silencing of very extensively paired mRNAs. Our cells actually don't have, most of our cells don't have endogenous siRNAs, but we do have these microRNA duplexes, and so the machinery is there to be able to use these duplexes to slice the mRNA. But in most cases in animals, it's not through this pathway. Instead, it's going in this other direction, where you have the, there's much less extensive pairing between the microRNA and the mRNA. And what's really key for this other mode of target recognition is pairing to what's called the seed of the microRNA, two, nucleotides 2 through 7. And that seed pairing is often supplemented by a pair here, to nucleotide 8, or an A here, across from nucleotide 1. And either of those will make a 7 nucleotide site, and both of them produce an 8 nucleotide site. And those are the types of sites that are most frequently conserved in the UTRs. It's actually surprising, there's very little pairing, or very little role for pairing to the 3', in the middle region of the microRNA. Sometimes, you have some pairing there that supplements the seed pairing, but actually not so often. Less than 5% of the conserved targets of these animal microRNAs have that 3' supplementary pairing. In other types, you'll enough, though, pairing here to the 3' end, that it can actually compensate for a mismatch, or wobble, in the seed region. But that's even more rare, where less than 1% of the conserved targets have that sort of 3' compensatory pairing. So, what I've told you so far about this type of pairing, actually, we've found, in, not from experiments, but actually looking using computational methods to look at the types, of all the different types of pairing to the conserved microRNAs, what are the types of pairing that are preferentially conserved in the UTRs. And the reason that that worked is that so many messenger RNAs were under selective pressure in mammals to preserve their pairing to the conserved microRNAs. So, the take-home here is that the conserved mammalian microRNAs have many conserved targets. If you, in our most recent analysis, if we're just focusing on the 87 conserved families that are shared between human and fish, some are even further, but at least those 87 that are conserved from human to fish, on average, for each of those conserved families, there are more than 400 human messenger RNAs that have been under selective pressure to maintain their pairing to the microRNAs. That's after we take out what you'd expect by chance. Now, often, when a microRNA has a conserved site to an mRNA, that mRNA has conserved sites to other microRNAs. On average about 4 to 5 conserved sites per conserved target. And so, often, these aren't necessarily two microRNAs that are always expressed at the same time, but you still get the idea that the microRNAs are working together to downregulate many of these targets. And even with hitting multiple conserved sites, to different microRNAs in the same targets, this conserved targeting here adds up to more than half of the human mRNAs. More than half the human messenger RNAs are conserved targets of microRNAs, over 60%. So, this was actually the second big shift in our thinking about microRNAs. The first was that the, well, that there are many microRNAs, and the second was that they have such widespread targeting in animals, humans and other animals. And really, this is the conserved targeting, and they're even, when you look experimentally, because these 7 nucleotide sites are very frequently, not always, but frequently, sufficient for mediating repression, there's actually more non-conserved targeting that could serve targeting. (??) So, if a biologist is interested in the regulation of a human gene, or a gene in other animals, changes are, at some point in development, that gene is going to be regulated by microRNAs and it's going to be very hard to find a disease or developmental process that isn't somehow influenced by microRNAs. So, how are we thinking about this now? Well, the messenger RNAs are regulated of course by chromatin and transcription, and you can have some messenger RNAs in some cells that are not expressed at all. Those promoters are turned off. Others will be expressed at very high levels and others, in other cells, it'll be intermediate levels. So, the same gene, different expression levels in different cells. And then over the course of evolution, the mRNA from that gene acquires sites to these microRNAs. And the more sites you have, the coexpressed microRNAs, the less protein output that you have in that cell type. And, of course, some of these sites will be more effective than others, other sites will be for microRNAs that are in some cells and not in other cells, and in this way, you can get very complex patterns of gene expression just starting with rather simple promoters for the mRNAs and simple promoters for the microRNAs. Ok, so this might be, although it seems wasteful, you know, at first glance, that you're making this mRNA just to later have it degraded, might actually be one of the more accessible ways to get these complex gene expression patterns that are needed for animals. So, you can think about this in another way, as transcription here as creating this column of gene expression and that microRNAs are the sculptor, the artist, that's chipping away at that gene expression, actually at the mRNA level, we know now that most of the effects are happening to decrease the amount of mRNA, so the mRNA is being destabilized, being chipped away by the set of microRNAs that pairs to each of those mRNAs. And in this way, for some targets, some key targets, that can actually promote a developmental transition, as was seen, originally for the lin-14/lin-4 interaction. In other cases, it can make that developmental transition much sharper, in, as is seen in early development, in the zebrafish. But more generally, what's happening here is that this chipping away at the gene expression produces a much more complex topology of gene expression, and a more optimal amount of protein in each of these different cell types. So, how does that chipping away at the gene expression actually take place? Well, here is the current understanding of the mechanism for microRNA repression in animals through these seed match sites. So, the seeded microRNA brings the silencing complex to the 3' UTR of the messenger RNA. And, sometimes you can also have effective sites in the open reading frame, but most of the time, it seems to be in the 3' UTR. The silencing complex then recruits this adapter protein, called GW182, that's the name in flies, and that GW182 protein interacts directly with this poly-A binding protein, which interacts with the poly-A tail of the mRNA, and at the same time, it recruits this complex of proteins called a deadenylase complex and the most important deadenylase complex here is the Ccr4-Not complex, which has a couple of deadenylase proteins, proteins, enzymes, that shorten the poly-A tail, and then once that poly-A tail gets short enough, then the mRNA is decapped and degraded from the 5' end. So, in this way, the microRNA is destabilizing the mRNA, through the same processes that are normally happening to mRNAs, but just accelerating them by recruiting this deadenylase complex. And the deadenylase complex is also thought to have a second role, in our experiments show that it doesn't have as much of an effect as the mRNA destabilization, but still, an important role in lowering the amount of protein that's made from these messages by inhibiting translation initiation. So, this is the mechanism, the current idea, of the mechanism for the microRNAs. Why is this repression important? Well, this has been looked at in thousands of different cases, I'll just present a few of them here. One approach that experimentalists do is they'll just look broadly at the importance of the microRNAs, they'll eliminate one of the enzymes that's needed to make the microRNAs, like Dicer, or Drosha. And so here's an example in zebrafish, where Alex Schier's lab has made fish that doesn't have any microRNAs, or not very many of them, because there's no Drosha, I'm sorry, no Dicer in these fish. So, here you have fish that really have almost no microRNAs, and you can see that the fish has very severe problems with the brain, and other parts of its development. It actually gets pretty far though, it can get the different cell types can be made, the different tissues and organs are made, but there are many, many problems without the microRNAs. In mice, actually, if you do this experiment, get rid of Dicer, and, this is what Greg Hannon's lab has done, they got rid of Dicer, and the mouse embryo doesn't make it very far at all, it dies very, very early in development, so you can't really do this type of experiment in mice, what people do in mice is they make conditional mutations of Dicer, just eliminate from certain sets of cell types, or organs. But in fish, you can get all the way to this stage, where you can see these brain defects, and then what's interesting here is that they were able to rescue a large part of this brain defect just by adding back one microRNA, already processed microRNA duplex for mir-340. So, this is a microRNA that's very highly expressed in embryonic development and this experiment shows that part of the reason that it's expressed so high in the embryo is that it's needed for proper brain development and a control, obviously does not give that same effect. Well, doing this Dicer experiment in humans, obviously is not something you'd want to do, but we can important clues about microRNA function just looking at cells that come from patients that have had the misfortune of having their microRNAs disregulated, and that has helped lead to certain types of cancers. So, an example of this is with the mir-17/92 cluster. This comes from a region of chromosome 13 that is very often amplified in certain types of lung cancer and lymphomas. So, that's a clue that it is somehow driving the formation of those tumors. And in fact, Greg Hannon's lab and their collaborators have shown that for mice that already have a propensity to have these lymphomas, that when you increase the expression of this cluster of microRNAs, that these mice get these lymphomas much more rapidly. So, here you have a set of microRNAs that really have all the important features of an oncogene but instead of it being, producing a protein, the oncogene here is this set of microRNAs. In other cases, too little of a microRNA can cause a problem. There's another case, right here again, on chromosome 13, where you have this pair of microRNAs that is very often deleted in certain leukemias, and other cancers. So, this set of microRNAs has the features of a tumor suppressor gene. And there are other cases like this. Another thing that can happen is that the regulation by the microRNA can be disrupted in cancers. So, these tumors will sometimes have what are called translocations where part of one chromosome is switched with part of another chromosome. So, in this case, this is just the painting that was done in the case where you had a translocation between chromosomes 3 and 12. And you can see that here you have a case where there's mostly chromosome 3, but attached to that is a little bit of chromosome 12, and then you have the reciprocal event here where 12 is attached to chromosome 3. So, this is a type of translocation that can happen in cancer. And there's an enrichment for these types of translocations at the HMGA2 locus. For this leukemia and for other types of tumors. Now, HMGA2 is an oncogene, too much of this will drive tumor formation, and so what's happening here is that when there's a translocation, the normal copy of HMGA2 is made, but there's also another copy that does not have the long 3' UTR that's normally seen in HMGA2. And that long 3' UTR has seven highly conserved sites to the microRNA let-7. That was the microRNA originally found by Gary Ruvkun, which turns out to be a tumor suppressor gene, in part because it is downregulating HMGA2. And so what happens in these tumors with the translocation, they've swapped out the 3' UTR, they have a slight change here in the open reading frame, that doesn't turn out to be what's important; what's important is that they no longer have the UTR that can be regulated by let-7. And for that reason, you have, this translocation can help promote these tumors. So, this is another example where microRNA regulation is very important in humans. So, and there are many, many other examples that have been reported, and many more that still remain to be discovered, so this is a very exciting area of research, I hope you've enjoyed this introduction about microRNAs and I hope that you'll stick around and take a look at the next two parts of this series, where I'll talk about some of the experiments that we've been doing to measure the effects of microRNAs, and to answer the question of, what is a microRNA and what isn't? So, thanks again for listening and have a great day.
