A Perfect Storm and Small Non-coding RNAs

   

                                      Overview

         The discovery of small non-coding RNAs that regulate mRNA expression and translation levels added an exciting new layer of complexity to the control of gene expression. In his talk, Gary Ruvkun describes the experiments that led to the identification of the first microRNA, lin-4, which downregulates the translation of lin-14, a protein needed in the early development of the model organism C. elegans. Since then, a wide variety of small regulatory non-coding RNAs that affect numerous cellular processes have been discovered in almost every organism.

          

                                              Transcript

00:00:08.06 the small RNA revolution that occurred
00:00:10.13 between 1990 and today, and how that emerged.
00:00:15.22 My name is Gary Ruvkun.
00:00:17.00 I'm a professor of genetics are Harvard Medical School.
00:00:20.15 One learns as an undergraduate,
00:00:22.12 the central dogma of molecular biology,
00:00:25.13 where DNA makes RNA, that makes proteins.
00:00:27.25 And small RNAs have modified that over the last 20 years.
00:00:32.25 The first example of a modification was the microRNAs,
00:00:36.09 that I will go into great detail today,
00:00:37.22 and they are thought to be made
00:00:39.11 from very small transcripts, 22 nucleotides long,
00:00:43.22 that then regulate the translation of target mRNAs,
00:00:47.28 to then cause less protein products.
00:00:51.00 So it's another layer of regulation.
00:00:52.28 But in addition to microRNAs, there are also
00:00:56.19 many other tiny RNAs that are made inside of cells
00:01:00.04 that probably regulate the production of RNA from genes
00:01:06.09 and also add a regulatory locus.
00:01:08.21 And those were all discovered at about the same time,
00:01:10.17 and I'll talk about that, too.
00:01:11.24 So the star of the show is C. elegans, the nematode.
00:01:15.22 And this is the animal that is not as big as my arm,
00:01:20.01 it is is a millimeter long and it has been
00:01:24.24 a subject of thousands of researchers worldwide who work on it.
00:01:29.20 It started as a field by Sydney Brenner in the early 60s,
00:01:34.21 as the most simple complicated animal to work on.
00:01:37.23 The worm is an animal and it's phylogenetically related
00:01:44.04 to us up here, but pretty distant.
00:01:47.29 And much simplified as an animal.
00:01:50.25 So simple that instead of having billions of cells,
00:01:55.13 it has 959 cells as a full grown adult.
00:01:58.19 This is the lineage of the animal,
00:02:00.24 it was worked out by John Sulston in the early 70s.
00:02:04.10 This is showing the pattern of divisions
00:02:06.21 that take place in the embryo.
00:02:08.05 These are the divisions that take place during larval stage 1,
00:02:11.15 as you can see here, larval stage 2,
00:02:14.16 larval stage 3, larval stage 4, and out pops an adult.
00:02:18.08 It produces progeny every 3 days,
00:02:21.15 so it's a very fast genetic system
00:02:24.14 that attracted a lot of talent.
00:02:25.27 And the field of C. elegans has been
00:02:28.23 very influential to me, in particular.
00:02:31.09 It's just populated with really wonderful people
00:02:35.07 who work on many different biological problems.
00:02:36.26 And I've always felt blessed to have been
00:02:39.20 surrounded by people who work on
00:02:43.06 so many different aspects of the biology of this animal.
00:02:46.05 It's a way to learn everything about biology
00:02:48.29 in a microcosm on one organism.
00:02:50.22 The particular project that I'll be describing in detail
00:02:56.21 was started at the MRC,
00:03:00.14 which is the center of Medical Research Council,
00:03:02.15 which was the center where worms started in Sydney Brenner's lab.
00:03:05.05 Actually Marty Chalfie started this project,
00:03:07.20 discovering the lin-4 mutation, back in the mid-70s.
00:03:14.12 And this is the mutation that causes lineage changes
00:03:18.02 and those lineage changes look like changes in time.
00:03:21.27 So if you look at the wildtype lineage over here,
00:03:24.22 this complicated pattern of divisions that takes place
00:03:28.10 in the larval stage 1, only takes place in larval stage 1
00:03:32.07 in wild type, yet in the lin-4, it goes at the L1, L2, L3, and L4 stage.
00:03:38.05 It keeps reiterating, so this animal is retarded
00:03:42.16 and never really gets to the adult stage.
00:03:45.00 It always has patterns of division as if
00:03:48.07 it's a larval animal.
00:03:49.27 And what really sort of launched this project
00:03:54.11 to be really important to do,
00:03:57.14 was when Victor Ambros began to work on it
00:04:01.08 when he was a postdoc in Bob Horvitz's lab in the early 80s,
00:04:05.23 and discovered that gene lin-14 could suppress
00:04:08.25 the mutations in lin-4.
00:04:10.22 So it could take this complicated set of divisions
00:04:13.06 that are temporally abhorrent,
00:04:16.23 and instead of doing L1, L1, L1,
00:04:20.06 it would go straight to L2,
00:04:21.28 and therefore have a pattern of divisions
00:04:25.23 that's distinct in time.
00:04:27.12 And this allowed him to infer that
00:04:31.08 lin-4 was the negative regulator of the gene lin-14.
00:04:35.00 To set up how this developmental timing works,
00:04:38.25 And so Victor and I worked together as postdocs on this project,
00:04:43.09 to try and figure out how lin-4 and lin-14 regulate
00:04:46.23 former regulatory axis to set up the timing of the animal.
00:04:51.20 So we together identified the genetic locus
00:04:57.14 the actual molecular basis of lin-14
00:05:01.14 and the first thing we figured out
00:05:04.02 was that mutations that activate the lin-14 gene
00:05:08.13 are deletions not of the open reading frame of the gene here,
00:05:12.08 it's downstream of the open reading frame,
00:05:14.00 but actually mutations in the 3' untranslated region
00:05:17.13 or translocation of the 3' untranslated region,
00:05:20.26 that takes these conserved regions shown in gray
00:05:23.09 and either deletes them or translocates them away.
00:05:28.09 So that said that there's an off-switch on the lin-14 gene
00:05:33.09 and this is part of the off-switch.
00:05:35.05 And the way it's turned off,
00:05:37.14 is by this negative regulator, lin-4, shown here.
00:05:41.27 And so what was lin-4?
00:05:44.04 The key breakthrough here came out of Victor's lab
00:05:46.13 when he was an assistant professor at Harvard,
00:05:49.27 and what his lab figured out was
00:05:51.23 Rosalind Lee, and Rhonda Feinbaum, and Victor
00:05:54.25 was that lin-4 didn't code for a protein at all,
00:05:57.27 it codes for very small RNA that's 22 nucleotides long
00:06:02.21 that's shown in the gray box here,
00:06:04.09 the precursor for lin-4 is shown as this double stranded RNA.
00:06:09.21 And this RNA is upregulated right at the time of L1
00:06:15.13 which fits with the phenotype of what we were studying.
00:06:19.01 So after the Ambros lab figured out that lin-4
00:06:22.13 encodes a tiny RNA, we, Victor and I,
00:06:24.27 compared the sequence of his lin-4,
00:06:28.24 shown on the bottom strand here,
00:06:30.22 to the sequence of lin-14,
00:06:32.10 assuming it might be a direct interaction,
00:06:35.23 which you never know if it's direct or indirect in biology,
00:06:38.06 but let's assume simplicity.
00:06:39.21 And so we could immediately see
00:06:42.22 that there were 7 different sites on the lin-14 UTR,
00:06:46.16 so they're numbered 1-7 here,
00:06:48.11 and for example, here's number 1,
00:06:51.18 and that's shown over here,
00:06:54.05 the top strand is lin-14, the bottom strand is lin-4.
00:06:58.19 And in every case, there was a duplex,
00:07:02.04 but as you notice, it's not a perfect duplex.
00:07:04.00 The lin-4 would bulge a C here,
00:07:06.00 it'll bulge an ACCUCA over here,
00:07:09.10 and each one had a sort of non-perfect duplex.
00:07:14.12 On the other hand, we could see by evolutionary comparisons,
00:07:18.09 comparing the sequence of lin-14 on the top strand
00:07:22.06 to between Caenorhabditis elegans and briggsae,
00:07:25.22 that all the bolded letters are conserved
00:07:28.27 between C. elegans and C. briggsae.
00:07:30.22 So that said the sequence of this is important enough
00:07:34.06 for the animal to have conserved it over many
00:07:36.13 billions of years.
00:07:37.19 And so we proved that these elements were important
00:07:43.02 by moving this 3' UTR into reporter genes
00:07:47.01 and knocking out these sites to prove it was important.
00:07:50.16 And this was the first example
00:07:55.14 of a regulatory RNA that was that small
00:07:59.05 in the case of lin-4.
00:08:00.18 And it was much smaller than any other regulatory RNA,
00:08:04.21 for example the uRNAs are more like 100 nucleotides,
00:08:08.00 than 20 nucleotides.
00:08:09.11 In this developmental biology community,
00:08:14.27 the community that we were a part of,
00:08:16.13 that included, for example, many C. elegans geneticists,
00:08:20.16 drosophila geneticists, they paid attention to this.
00:08:24.01 We definitely talked about this at Gordon conferences
00:08:26.12 and things like that.
00:08:27.10 But there was always a view that this was
00:08:29.20 kind of a quirky thing, that may be a quirk
00:08:33.08 of the lineage of C. elegans.
00:08:35.11 Because remember, we're dealing with a lineage here
00:08:39.15 which was a distinct way to think about developmental biology.
00:08:43.12 And really sort of marginal in comparison
00:08:48.05 to how the wider developmental biology community thought
00:08:51.05 more about tissue interactions and cell interactions.
00:08:54.10 And so, yes, we could present it at meetings and things like that,
00:08:58.24 but it wasn't really, there wasn't a stampede
00:09:01.15 for people to start work on it, that's for sure.
00:09:03.25 On the other hand, the regulatory RNA community
00:09:07.05 the people who work on the ribosome,
00:09:09.02 how the tRNAs bind acceptor sites,
00:09:11.03 that sort of thing, splicing,
00:09:13.13 they were totally enthralled.
00:09:15.14 So it was exciting to us,
00:09:19.07 to be thrust into the regulatory RNA community.
00:09:23.06 Remember, we're coming as developmental biologists,
00:09:25.13 what we happened to find was a regulatory RNA
00:09:29.15 and it's target, thrust us into a new ecosystem.
00:09:33.01 And that ecosystem informed us in many ways,
00:09:36.27 that it's very sophisticated.
00:09:39.02 How the ribosomal RNA folds,
00:09:41.08 how the ribosome works,
00:09:43.06 these are inspirational fields.
00:09:46.08 And it's one of the great things about biology,
00:09:49.24 there are these islands of specialization,
00:09:52.29 and you can learn a lot from each island.
00:09:58.18 Biology teaches us that the islands are where
00:10:03.01 the most extreme variation occurs.
00:10:06.04 So how does the lin-4 microRNA work?
00:10:10.27 Well, lin-4 gets upregulated in the L1 stage
00:10:14.23 and that causes less production of the lin-14 protein,
00:10:17.27 that's its target, but it doesn't really affect the mRNA levels
00:10:21.22 nearly as much as it affects the protein levels.
00:10:23.26 So that said that the way that microRNAs work
00:10:26.09 is that one microRNA is by regulating the transcription of target proteins.
00:10:31.03 We still don't know, even 20 years later,
00:10:33.24 exactly how that's working.
00:10:35.16 It's still an object of much research in microRNAs.
00:10:41.22 The next stage, right now we're at 1993,
00:10:47.13 approximately the next stage is 7 years later,
00:10:50.07 when we start doing genetics to try and fill out the rest
00:10:53.28 of the pathway. To figure out who else is being regulated.
00:10:57.03 To specify these lineages, and we get let-7
00:10:59.17 as a suppressor of a weak lin-14 mutation.
00:11:03.06 And when we figure out what let-7 is,
00:11:05.29 we figure out that it's another microRNA.
00:11:09.04 In this case, looks a lot like the first one,
00:11:11.29 it's not homologous at all, but it's analogous.
00:11:14.00 It has the same stem loop structure,
00:11:15.21 the gray box is the 22 nucleotide product of that.
00:11:19.21 And in this case, the let-7 microRNA,
00:11:24.24 when we compared it's sequence across emerging databases,
00:11:28.27 now remember, this is the year 2000,
00:11:30.23 as opposed to 1992, so genome databases
00:11:33.20 are much more sophisticated.
00:11:35.22 The human genome in the year 2000 is about 30% done,
00:11:39.09 so we can compare it to the 30% done genome sequence,
00:11:42.22 not by doing northern blots of anything like that,
00:11:45.10 but by just doing a BLASTn analysis,
00:11:48.00 just asking "is the nucleotide conserved?"
00:11:50.10 And boom! We could see it 20 seconds later.
00:11:53.11 And that was an amazing moment,
00:11:56.10 because the activation energy to actually do
00:12:00.15 a northern blot or some kind of experiment
00:12:03.00 to look for conservation takes some oomph to do that.
00:12:06.15 But looking at a genome, you could do it in 20 seconds
00:12:09.11 and know the answer.
00:12:10.10 And so we sort of instantly knew that let-7 was conserved
00:12:14.13 in flies, in humans, in genome sequences,
00:12:18.17 but we also had to of course prove that it was a small RNA.
00:12:21.22 So we did do a lot of northern blots
00:12:23.11 to a whole zoo of creatures.
00:12:25.11 And that was very influenced by comparative
00:12:31.28 developmental biology conferences,
00:12:35.08 that Eric Davidson had run at the Marine Biological Labs,
00:12:38.16 which is where I'm actually speaking to this camera right now.
00:12:41.08 Now at about the same time,
00:12:44.06 that we were finding this second microRNA,
00:12:46.06 in the year 2000.
00:12:47.23 Fire and Mello, and Baulcombe and Hamilton,
00:12:51.27 had found that tiny RNAs are produced in RNA interference,
00:12:57.19 which was a very hot topic at that point.
00:13:00.20 It was very mysterious, how injecting double stranded RNAs
00:13:04.02 into organisms would inactivate genes.
00:13:05.25 Hamilton and Baulcombe found 22 nucleotide RNAs.
00:13:09.17 I was enthralled by the fact that the microRNAs
00:13:14.20 and the siRNAs are the same size.
00:13:18.16 I remember the number 22 was an important number
00:13:22.22 in the Kabbalistic numerology,
00:13:25.09 and sort of as a joke,
00:13:26.27 I found websites to show my lab,
00:13:28.28 and they worried about me doing this.
00:13:31.03 Much more productive was to say,
00:13:34.27 okay, if they produce the same size RNAs,
00:13:39.17 maybe they use some of the same enzymology.
00:13:41.16 And the enzymes that take double stranded RNAs
00:13:45.24 and chop them into siRNAs were being discovered.
00:13:48.10 For example, by Greg Hannon's lab at Cold Spring Harbor,
00:13:50.28 and so we could do the experiment of inactivating
00:13:54.09 the gene for that double-stranded RNA, Dicer,
00:13:58.11 RNase Dicer, and show that was actually important
00:14:01.11 in how microRNAs get processed as well.
00:14:04.18 The other importance of the synchrony
00:14:09.13 of the discoveries of the RNAi pathway and the micro pathway
00:14:13.01 is that there was a huge interest in RNAi as a tool.
00:14:18.15 The C. elegans community was probably one of the very first
00:14:22.13 genetic communities to start using RNAi in a big way.
00:14:25.20 But it also started to happen a lot in plants,
00:14:28.21 and once siRNAs were discovered in animals,
00:14:32.03 they started to be used all over animal labs.
00:14:35.04 So the interest in it as a tool was immense,
00:14:39.17 and it made the interest in microRNAs
00:14:43.17 and small RNAs in general, much larger
00:14:46.02 than if it was just microRNAs.
00:14:48.09 So after let-7 was discovered as a conserved microRNA,
00:14:53.13 there was a concerted effort to purify by cloning
00:14:57.14 many different microRNAs.
00:14:59.06 And actually, Victor Ambros's lab was one
00:15:02.02 of the first to do that, along with the Bartel lab
00:15:04.14 and the Tuschl lab.
00:15:05.11 And many microRNAs were discovered,
00:15:07.26 and within a year of two, there were thousands
00:15:09.24 of microRNAs, the current census must be
00:15:13.09 tens of thousands of microRNAs.
00:15:15.06 And it is known now that microRNAs regulate many many
00:15:19.27 different targets, and are regulating those targets
00:15:23.13 to regulate all kinds of processes.
00:15:26.20 They work everywhere from floral development
00:15:29.04 in plants to perhaps how memories are formed.
00:15:31.25 There are thousands and thousands of papers on microRNAs
00:15:36.07 in all different fields, from plants and animals.
00:15:39.21 More generally, the small RNA field,
00:15:43.21 it's thought that there are many sources of
00:15:47.01 double-stranded RNAs, one of those sources is
00:15:50.18 microRNAs that I started with,
00:15:52.08 but there's also viral double-stranded RNAs,
00:15:55.03 repeats in the genome,
00:15:56.17 they go through this process of making tiny RNAs
00:16:00.13 using Dicer and various other components
00:16:03.05 and they then regulate gene expression.
00:16:05.13 They regulate gene expression by a variety of mechanisms,
00:16:10.03 the ones I've told you about so far
00:16:12.10 are translational repression and RNA degradation,
00:16:16.22 which siRNAs mediate by simply causing cleavage.
00:16:19.18 More mysterious is the regulation of heterochromatin
00:16:23.23 which also involves small RNAs,
00:16:25.10 and how that histone modification response
00:16:29.17 to small RNAs is a big mystery today.
00:16:31.14 Thank you very much.
00:16:34.07