X Chromosome Inactivation • iBiology (2024)

00:00:15.06Hello, everyone.
00:00:16.18I'm Jeanie Lee.
00:00:17.28I'm a Professor of Genetics at Harvard Medical School,
00:00:20.03and I'm also a faculty member
00:00:21.28within the Department of Molecular Biology
00:00:23.23at Massachusetts General Hospital.
00:00:25.14What I'd like to tell you about today
00:00:27.15is a scientific problem that we've been working on
00:00:29.27for the past 22 years or so,
00:00:32.21and that is the problem of why and how
00:00:36.21one of our sex chromosomes needs to be inactivated
00:00:39.29in early development.
00:00:41.14So, this is a process called X chromosome inactivation.
00:00:44.26And what I'm going to do is divide this topic
00:00:48.14into three lectures.
00:00:50.07So, in the beginning we'll talk about
00:00:52.22a general overview of the scientific problem
00:00:55.04and the medical relevance of X chromosome inactivation.
00:00:58.26And then we'll take a deeper dive,
00:01:01.05in the second and third lectures,
00:01:02.25and talk about how inactivation initiates
00:01:05.20with a counting and choosing mechanism.
00:01:08.01And then end with a very challenging problem
00:01:13.15of how X inactivation initiates within a tight locus
00:01:18.23before spreading to the rest of the X chromosome.
00:01:21.17Okay.
00:01:23.04So, we'll start with...
00:01:24.26by taking a step back and asking this question of,
00:01:27.03why do we need to inactivate
00:01:29.06one of our sex chromosomes?
00:01:30.17And I think the answer lies in the fact that,
00:01:33.05in mammals, sex is determined by a pair
00:01:35.22of the so-called unequal sex chromosomes.
00:01:38.18And so, here what you see is
00:01:41.07a normal human female karyotype,
00:01:43.22so it's 46 XX.
00:01:46.00And you'll notice that these chromosomes
00:01:49.04come in pairs.
00:01:50.21And so humans have 23 pairs of chromosomes.
00:01:53.22And the first 22 are so-called autosomes.
00:01:59.00And we obtain one copy of each autosome from mother
00:02:04.14and one copy from our father.
00:02:06.16Okay?
00:02:08.04And for all intents and purposes, for this lecture,
00:02:10.11we'll say that the two chromosomes that we get from mother and father
00:02:13.16are genetically identical.
00:02:15.22Now, the 23rd pair is very special,
00:02:19.18and that's because these are our sex chromosomes.
00:02:22.17So, we call them sexually dimorphic
00:02:24.27because they're different between males and females.
00:02:28.07And if we obtain one X chromosome from our mother
00:02:31.14and one from our father,
00:02:33.08we develop as genetic females,
00:02:35.22so 46 XX.
00:02:37.17But on the other hand,
00:02:39.14if we obtain one X chromosome from our mother
00:02:42.13and one Y chromosome from our father,
00:02:45.24we develop as genetic males.
00:02:48.12Okay.
00:02:49.13So, to simplify this scheme,
00:02:51.00here I've drawn two X chromosomes belonging to the female,
00:02:54.24and to the male one X and one Y chromosome.
00:02:59.01So, the X chromosome is a very large chromosome.
00:03:02.16It accounts for about 5% of our genome,
00:03:04.21and it has approximately 1000 genes.
00:03:08.13By contrast, the Y chromosome is much, much smaller
00:03:12.05and only contains a fraction of genes.
00:03:15.18And so this mechanism of determining sex in mammals
00:03:20.06produces this genetic inequality between the sexes
00:03:23.19that needs to be compensated
00:03:26.10in order for development to proceed normally.
00:03:29.16And so, to compensate,
00:03:31.03we mammals do something truly unique,
00:03:33.24and that is a process called X chromosome inactivation.
00:03:37.08And what we literally do in early female development
00:03:40.25is to transcriptionally silence one of the X chromosomes
00:03:44.25in the female sex.
00:03:46.17And so what we have, then,
00:03:48.10is an epigenetic situation
00:03:51.09in which males and females
00:03:54.29express only one functioning copy of the X chromosome.
00:03:59.20So, this is very interesting,
00:04:01.07because the X chromosome is in fact
00:04:04.06the only chromosome in mammals
00:04:06.20that's capable of undergoing this global inactivation.
00:04:11.08And so the mechanisms that underlie this process
00:04:14.28have been intensively investigated for more than half a century,
00:04:19.04firstly because it is a very nice model
00:04:22.02by which we can study something called non-coding RNA,
00:04:25.05and also to study general mechanisms
00:04:27.05of epigenetic silencing.
00:04:29.20And the X chromosome, of course,
00:04:31.29is also very interesting to both basic scientists and to clinicians,
00:04:35.18because of the numerous disease genes
00:04:38.19that are associated with the chromosome.
00:04:40.27So, I'm gonna tell you more about these things
00:04:43.11in a few minutes.
00:04:45.01But first, I wanted to just say a few words
00:04:47.13about the short history of X inactivation.
00:04:50.03So, the inactive X chromosome
00:04:53.16was actually first visualized
00:04:56.06by a pair of Canadian scientists, Barr and Bertram.
00:04:59.06So, in 1949, what they published was that
00:05:04.20they could, in almost all female nuclei,
00:05:07.15observe this structure,
00:05:10.06which you might be able to see here as a small dot
00:05:13.14lying next to the nucleolus.
00:05:15.19Okay.
00:05:17.01Whereas that structure is not present in these male nuclei.
00:05:21.18Now, they didn't actually know what they were observing at the time.
00:05:26.15They didn't know that this was an inactivated X chromosome.
00:05:30.04But what they did mention was that
00:05:32.24even without looking at these cats,
00:05:35.04which is...
00:05:37.06these were the organisms that they were working with...
00:05:39.02without laying their eyes on the cats,
00:05:41.28they could predict the sex of that animal
00:05:45.06simply by the presence or absence
00:05:49.18of the so-called Barr body,
00:05:51.15which has been named in honor of the scientist
00:05:53.24who first visualized the structure.
00:05:57.04Okay.
00:05:58.23So, fast-forward 10 years.
00:06:00.10An American scientist, Susumu Ohno,
00:06:02.07took this observation one step further.
00:06:05.06Now, he was looking in rats,
00:06:08.00but rats are also mammals...
00:06:09.17and we do some...
00:06:11.10and mammals do a similar process of X chromosome inactivation,
00:06:15.02and what he observed is that
00:06:18.09the number of Barr bodies in the rat
00:06:20.27very closely parallels
00:06:24.00the number of X chromosomes in that cell.
00:06:27.11And he called the Barr body an X chromosome
00:06:32.07and suggested that it was pyknotic,
00:06:35.15which is a fancy word for condensed,
00:06:38.13and potentially biochemically distinct,
00:06:41.21and maybe even functionally disabled.
00:06:46.00Okay.
00:06:48.10So, then... the person who really put all of this together
00:06:51.03was a British scientist, Mary Lyon, in 1961,
00:06:54.15who published a seminal paper in the field,
00:06:57.02in which she calls the Barr body
00:07:00.01an inactivated X chromosome,
00:07:02.12and suggests that X chromosome inactivation
00:07:05.26is the mechanism of dosage compensation
00:07:08.17that allows males and females
00:07:11.01to equalize their sex chromosome dosage,
00:07:14.04and that it would lead to a random inactivation
00:07:18.01of one X chromosome in the female cell.
00:07:21.17Now, she deduced all of this
00:07:23.24by working in mice and looking at coat color variegation,
00:07:27.10but the same phenomenon holds true in these cats.
00:07:30.16Okay?
00:07:32.02So, in cats and in mice,
00:07:36.06the coat color gene is carried on the X chromosome.
00:07:40.20And it comes in two flavors:
00:07:42.04there's an orange flavor, and then there is a gray flavor.
00:07:45.19Alright?
00:07:47.01So, what she noticed in mice
00:07:49.01-- but we're illustrating here with cats --
00:07:50.28is that while males can adopt
00:07:54.26either a solid orange color or a solid gray color,
00:07:59.08depending on whether the male inherited an orange...
00:08:04.21here's the... here's an orange coat color gene
00:08:07.00or a gray coat color gene...
00:08:09.15females can not only be solid orange or solid gray,
00:08:15.26but could have this very interesting pattern of variegation
00:08:19.25that you can see in this beautiful calico cat, okay?,
00:08:22.19if she inherited one copy of the gray
00:08:25.28and one copy of the orange gene
00:08:28.04on the X chromosomes.
00:08:29.25So, the idea here is that if a cell
00:08:33.18in this heterozygous female
00:08:35.19inactivated the orange-color chromosome,
00:08:38.16then that cell would manifest as a gray cell.
00:08:42.05But if a cell lying next to that gray cell
00:08:47.15chose instead to inactivate the gray gene,
00:08:50.11or the gray chromosome,
00:08:52.07then that cell would manifest as an orange cell.
00:08:55.22And so, in this manner,
00:08:57.19you can see this wonderful coat color variegation
00:09:00.00in the female calico cat.
00:09:02.14Okay.
00:09:03.19So, the important point to make here is that
00:09:05.16X inactivation is random.
00:09:07.12This is a choice step, which is random,
00:09:10.02and that is made by every single cell in the female body,
00:09:14.28and therefore every female mammal
00:09:17.19is a mosaic of cells that either
00:09:20.25inactivate the paternal chromosome
00:09:22.29or the maternal chromosome.
00:09:26.07Okay.
00:09:27.13So, we and the X inactivation field like to think about X inactivation
00:09:30.20as something that happens once,
00:09:32.14early in female development.
00:09:34.20But in fact, X inactivation happens in both sexes.
00:09:38.08And it happens as part of a much more complex
00:09:42.18life cycle of the X chromosome.
00:09:45.02So, here we are looking at
00:09:47.27the life cycle of X inactivation in the mouse,
00:09:49.13and we're going to start in the germline, here,
00:09:52.00with the female germline.
00:09:53.19You see that the two X chromosomes are active,
00:09:56.14and then one of the two X chromosomes
00:09:58.14gets passed on to the next generation,
00:10:01.18to daughter and to sons.
00:10:05.21Okay.
00:10:07.12So, now a very interesting process happens
00:10:10.05in the male germline, shown up here.
00:10:12.06So, during the first stage of male meiosis
00:10:16.19-- so, this is during spermatogenesis --
00:10:19.05the X and Y chromosomes physically pair,
00:10:22.04at least partially,
00:10:24.03and then they undergo a sex chromosome inactivation,
00:10:27.19of both the X and the Y chromosome.
00:10:30.16And then... the...
00:10:33.24we believe that that silencing remains in effect
00:10:37.29until the... until the end of spermatogenesis,
00:10:41.07at which time either the X or the Y chromosome
00:10:43.12gets passed on to the next generation.
00:10:45.27Okay.
00:10:47.10So, early pioneers in this field
00:10:50.06proposed the very interesting hypothesis
00:10:52.13that the X and Y chromosomes
00:10:54.28may be getting inactivated in the germline
00:10:57.10not only to suppress hom*ologous recombination,
00:11:00.25which would be detrimental to the sex chromosomes,
00:11:04.13especially if we wanted to preserve the identities of the sex chromosomes,
00:11:08.04but that it is also a mechanism
00:11:11.15by which the sex chromosomes could be imprinted.
00:11:14.23Okay?
00:11:16.13And be predisposed to silencing in the next generation,
00:11:19.02at least in daughters.
00:11:21.03So, we like that idea.
00:11:23.06And in fact, in the very first stages of development
00:11:28.21in the next generation,
00:11:30.18what we observe is an imprinted form of X inactivation,
00:11:34.24in which the father's X chromosome
00:11:37.09is always inactivated.
00:11:39.00Okay.
00:11:40.22And that persists until the blastocyst stage.
00:11:43.12You see here in the...
00:11:44.26this is a peri-implantation embryo,
00:11:46.21where the extraembryonic tissues will maintain...
00:11:50.26so, these are the placental...
00:11:52.20the tissues that will go to make the placenta.
00:11:54.24Okay.
00:11:56.08They maintain this pattern of paternal X inactivation.
00:12:00.05And the placenta will keep that until the time of birth,
00:12:04.27at which point of course the placenta is discarded.
00:12:07.26Now, on the other hand,
00:12:10.14in the embryo proper...
00:12:12.16so, we call it the epiblast lineage,
00:12:14.18and that is the lineage
00:12:16.17which will go to make all of the somatic cells in our body.
00:12:19.03Now, in the epiblast lineage,
00:12:21.07which is shown here,
00:12:23.02inside the blastocyst,
00:12:25.04the cells will erase the paternal imprint
00:12:29.03and undergo a reactivation
00:12:32.22so that the embryo itself can decide who...
00:12:38.00which of the two X chromosomes
00:12:40.04will become the inactive chromosome.
00:12:42.03So, there's a reactivation of the paternal X,
00:12:44.17and then a re-inactivation,
00:12:47.26except that this time inactivation proceeds in a random fashion,
00:12:52.25where either of the maternal or paternal chromosome
00:12:55.09could be inactivated.
00:12:58.19And that is the form of X inactivation
00:13:01.20that I'll be talking about today,
00:13:03.14and it is the form which persists
00:13:06.06throughout the rest of development
00:13:08.04and throughout the rest of adult female life.
00:13:11.20Okay.
00:13:13.06So, I'll briefly mention that in the germline
00:13:15.18of a female embryo,
00:13:18.06the two X chromosomes...
00:13:19.25well, of course, one is active and the other one is inactive...
00:13:22.12the inactive one will also undergo
00:13:25.05a reactivation event,
00:13:27.02so that in female meiosis the two X chromosomes
00:13:30.04would have an equal chance of being passed on
00:13:32.19to the next generation.
00:13:34.16And then of course the whole cycle reinitiates again.
00:13:37.18Okay.
00:13:39.14So, that is the life cycle of X inactivation
00:13:41.27in the mouse.
00:13:43.17And now we're going to address the question of
00:13:46.10why the X chromosome and X inactivation
00:13:49.14might be important to study.
00:13:51.03Okay.
00:13:52.18So, firstly, the X chromosome, as I mentioned before,
00:13:56.13is a very large chromosome,
00:13:57.28and it carries something like 1000 genes
00:14:00.22and makes up about 5% of our genome.
00:14:03.19And equally importantly,
00:14:06.08along with the Y chromosome,
00:14:08.04the X chromosome is responsible for
00:14:11.11so-called sexual dimorphism.
00:14:12.25So, that's a fancy term to say that
00:14:15.15males and females look different,
00:14:17.29and that we might even behave differently
00:14:20.01as a result of our sex chromosomes.
00:14:22.17Okay.
00:14:23.23And the X chromosome is enriched for genes
00:14:27.15that not only go to determine reproduction
00:14:30.17but also genes that influence
00:14:34.23brain development, behavior, and cognition.
00:14:39.16So, in fact, on the X chromosome
00:14:41.10there are more than 200 disease genes
00:14:44.12that are presently known.
00:14:46.01And many of these genes
00:14:47.26are responsible for neurodevelopmental disorders, autism, and other X-linked intellectual disabilities.
00:14:56.17Here I just list two known examples,
00:15:00.12so, Rett syndrome, which mostly affects girls,
00:15:02.15and Fragile X syndrome,
00:15:04.15which can affect both men and women.
00:15:07.24And then there's a muscular dystrophy that's fairly common,
00:15:12.00called duch*enne's muscular dystrophy,
00:15:13.14which mostly affects men.
00:15:15.06And then there are even non-disease traits,
00:15:17.23like red-green color blindness,
00:15:19.28which affects about 10% of all men
00:15:24.11across the world.
00:15:26.25Okay.
00:15:29.00So, then... you might also ask,
00:15:32.03what happens if X inactivation
00:15:34.09doesn't actually happen during development, right?
00:15:36.27So, this experiment was done
00:15:39.06more than 20 years ago in Rudolf Jaenisch's lab,
00:15:42.03in which they removed one of the critical factors for X inactivation
00:15:45.19called Xist,
00:15:47.05and observed that no female embryos were born.
00:15:49.22In fact, they all perished
00:15:51.18shortly after the time of implantation.
00:15:53.29Now then, we followed up many years later
00:15:56.24and asked, what would happen if we removed this critical factor
00:16:00.00a few days after conception?
00:16:02.29So, the original experiments
00:16:05.08were done from the time of conception.
00:16:06.15We asked, what would happen if we removed it several days later?
00:16:08.19And again,
00:16:10.17most of the female embryos perish.
00:16:12.19Okay?
00:16:14.22But a few make it to birth.
00:16:15.23But I think as you can see here,
00:16:17.23in the red circle,
00:16:19.16these female animals are almost always runted,
00:16:23.09growth-retarded,
00:16:25.01and do not make it past the third week of life.
00:16:27.11Okay.
00:16:28.28So, X inactivation is very important
00:16:30.29during early development.
00:16:32.15You might also ask the question,
00:16:34.06well, what if we were to take away X inactivation
00:16:36.26a little bit later in development,
00:16:38.23or even in the adult female...
00:16:41.14adult female cells?
00:16:43.13Okay?
00:16:45.06So, this has been a question that's been under
00:16:47.15intensive investigation for a number of reasons
00:16:50.05that will become obvious later.
00:16:51.28But the answer seems to be that
00:16:54.15it depends on what cells
00:16:56.16and when we do it.
00:16:57.27Okay?
00:16:59.02So, for example, in the mouse,
00:17:01.12in highly proliferative cells
00:17:03.26-- so, cells that are destined to divide many, many times,
00:17:07.04like blood,
00:17:08.28over the lifetime of a female --
00:17:10.16there seems to be an increased risk of cancer.
00:17:13.26But the same is not true in the brain,
00:17:16.08which... for the most part,
00:17:18.17once the brain is formed,
00:17:20.18does not... the cells within the brain do not divide again.
00:17:23.24Okay.
00:17:25.23So... and then in humans,
00:17:27.29there's also been a link to various cancers,
00:17:30.29although not currently shown to be...
00:17:36.09shown to be a direct causality.
00:17:39.00But what has been observed
00:17:41.24is that there is an increased risk of certain conditions,
00:17:46.24like blood cancer, breast cancer,
00:17:49.16ovarian cancer,
00:17:51.00and even, in men, testicular germ cell tumors.
00:17:54.16Okay.
00:17:56.08So, this increased risk is associated
00:17:59.18with both men and women.
00:18:01.04And they seem to be associated
00:18:05.10with the loss of the Barr body
00:18:07.23and/or the gain of additional active X chromosomes,
00:18:13.11suggesting that there's an increased risk of cancer
00:18:19.12when the X chromosome dosage
00:18:21.18exceeds physiological levels in both men and women.
00:18:27.15Alright.
00:18:29.10So, here's an experiment that we performed
00:18:31.22not so long ago
00:18:33.17in which we removed this critical factor, Xist,
00:18:35.25and demonstrated that these female animals
00:18:41.00-- and this was specific to the female animals --
00:18:43.13developed a fulminant blood cancer
00:18:46.00with essentially 100% penetrance,
00:18:48.19meaning that nearly all females
00:18:52.05succumbed to this disorder.
00:18:53.23And their blood stem cells become quote "cancerous".
00:18:56.18Okay.
00:18:58.01So, here's an example of one female that succumbed to the disease.
00:19:00.16And you can see that her spleen
00:19:02.23has become extremely large and filled with tumor cells.
00:19:07.14And just looking at the animal grossly,
00:19:10.19you can see that there are
00:19:13.24a number of abdominal masses,
00:19:15.15or masses that are present elsewhere in the body,
00:19:17.22indicative of a solid hematopoietic tumor.
00:19:22.09And if we follow these animals over their lifespan
00:19:26.09-- here I'm showing you a survival curve --
00:19:28.18you can see that most of the females
00:19:31.15perish in early to mid-adulthood.
00:19:34.28Okay.
00:19:36.22Alright.
00:19:38.18So, I think I've shown you that the X chromosome
00:19:43.04and its dosage compensation
00:19:45.13are very important,
00:19:47.15not only during embryogenesis
00:19:49.17but also throughout female life.
00:19:51.19And what I'd now like to turn your attention to
00:19:54.11is how we think the mechanism of X chromosome inactivation
00:19:58.07might be working.
00:20:00.15Okay.
00:20:01.27So, now, early pioneers in this field
00:20:04.08knew that there had to be a control center,
00:20:06.18and that that control center is most likely
00:20:09.22on the X chromosome,
00:20:11.14based on a number of genetic studies
00:20:13.13that we don't have time to get into right now.
00:20:15.23But there was a debate in the field
00:20:19.02about whether there was one inactivation center,
00:20:22.11located somewhere in the X chromosome,
00:20:24.15or a series of inactivation centers
00:20:26.12that would be located up and down the X chromosome,
00:20:28.24for silencing to spread.
00:20:31.14So, the experiments that were performed
00:20:34.08in Hunt Willard's lab,
00:20:36.19and also in my lab and the Jaenisch lab,
00:20:38.16in the 1990s
00:20:40.19showed that the inactivation center is in fact singular
00:20:44.26-- there's only one --
00:20:46.20and that that inactivation center is very small.
00:20:49.24It's probably no more than 100-200 kilobases.
00:20:53.22So, we're talking about a control center,
00:20:57.14a brain of the X chromosome,
00:20:59.04which is just 1/1000th of the size of that sex chromosome.
00:21:04.18And we were able to show that this inactivation center
00:21:08.01is indeed extremely powerful,
00:21:10.04because if we transplanted that inactivation center
00:21:12.24to a non-sex chromosome,
00:21:14.23to an autosome,
00:21:16.04that center will induce the autosome
00:21:19.03to behave like the sex chromosome
00:21:21.04-- be counted and be chosen for inactivation.
00:21:24.05In fact, the chromosome will undergo inactivation.
00:21:27.22And so what these genetic experiments told us
00:21:30.12is that X inactivation
00:21:32.29is driven entirely by this master switch
00:21:37.11that we call the X inactivation center,
00:21:39.12and that apart from this 200 kilobase sequence
00:21:42.28there are no...
00:21:45.04probably no other inactive X-specific elements
00:21:48.00that drive the process of silencing.
00:21:53.12Okay.
00:21:54.26So then, let's talk about the X inactivation center.
00:21:56.25So, here it is in all its glory.
00:21:59.02And you'll see that it's populated by an epigenetic...
00:22:02.17a type of epigenetic factor called
00:22:05.26a long non-coding RNA.
00:22:07.20So, I mentioned earlier that this region
00:22:10.15is probably no more than 200 kilobases in size.
00:22:12.29It's very small.
00:22:14.14And yet it is both necessary and sufficient
00:22:16.23to drive the different steps of chromosome silencing.
00:22:21.03Now, in the 1990s,
00:22:23.19sequencing experiments around this region demonstrated
00:22:27.04-- strangely --
00:22:29.06that there were few if any protein-coding genes.
00:22:32.29And that was very surprising for the 1990s
00:22:35.28because back then we all thought that anything that was important
00:22:39.11had to be encoded by a protein.
00:22:42.24So, in other words,
00:22:44.11we thought that proteins had to be the work...
00:22:46.28the workhorse for anything related to regulating gene expression.
00:22:52.01However, many of us felt that
00:22:54.20that may be an overly simplistic view,
00:22:56.17because in fact proteins make up only about 2% of our genome.
00:23:00.19The rest of the genome was termed, decades ago,
00:23:04.07as "junk DNA".
00:23:07.23They account for 98... 98% of the mammalian genome.
00:23:12.19And what's become very obvious in the past decade or two
00:23:16.04is that this so-called junk is extremely active.
00:23:20.10Nearly every single nucleotide of this junk DNA
00:23:22.27is synthesized into what's known as
00:23:26.22non-coding RNA,
00:23:28.16long non-coding RNA.
00:23:30.07So, the X inactivation center
00:23:33.02is a really good example of
00:23:35.25what long non-coding RNAs can do.
00:23:38.03In fact, it's enriched for this type of gene.
00:23:40.14So, I'll just mention a few here.
00:23:42.16So, there's the Xist gene,
00:23:44.21which was first identified by Willard and Ballabio.
00:23:48.15And that gene is responsible for
00:23:52.13spreading across the X chromosome
00:23:54.05and ushering in this silent state of the chromosome.
00:23:58.08And then we identified its antisense repressor,
00:24:01.03Tsix,
00:24:02.22back in the 1990s.
00:24:04.10This is an antisense transcript
00:24:06.13that prevents Xist
00:24:08.21from doing what it's supposed to do on the inactive X.
00:24:11.19And then there's a mysterious motif here,
00:24:14.21the repeat A motif,
00:24:16.05which is important for the transcriptional activation
00:24:19.17of the Xist gene,
00:24:21.24and also for recruiting polycomb complexes.
00:24:25.29And then, on the other side of Xist,
00:24:27.24we have two additional non-coding RNAs
00:24:30.07-- one called Jpx and the other called Ftx --
00:24:34.05both of which appear to be associated
00:24:37.14with the activation of this important Xist gene.
00:24:40.27Okay.
00:24:42.16So then, on the other side of Tsix,
00:24:44.03we have two other non-coding RNAs
00:24:46.00-- a Tsx gene and an enhancer element called Xite,
00:24:51.07which is responsible for
00:24:54.12inducing or allowing the expression
00:24:57.25of the antisense gene to persist.
00:25:01.28Okay.
00:25:03.13So, what we have here is
00:25:05.24a bipartite structure of the X inactivation center,
00:25:09.14in which on the left side
00:25:11.24we have all the pro-inactivation genes.
00:25:15.00These are genes that are necessary
00:25:18.12to induce the silent state of the X chromosome.
00:25:21.20And then on the other side,
00:25:23.06we have genes that are actively trying to prevent that
00:25:27.14from happening,
00:25:28.17so we call these the anti-inactivation genes.
00:25:31.25Okay.
00:25:33.09So, X inactivation is mechanistically complex,
00:25:36.03and it takes place in a number of different steps.
00:25:39.19Okay.
00:25:41.00So, we're just gonna go through these steps right now,
00:25:42.22and then I'll do a deeper dive in the next lecture.
00:25:45.09So, in the beginning,
00:25:47.01there is a counting mechanism,
00:25:49.11which is determining the number of X chromosomes
00:25:53.04in the... in the cell,
00:25:55.23and whether X inactivation should take place.
00:25:57.10And if the answer is yes, there is an allelic choice mechanism,
00:25:59.25which will allow the embryo
00:26:02.17to pick which of the two X chromosomes
00:26:05.14-- the mother's X chromosome or the father's X chromosome --
00:26:07.16for inactivation.
00:26:09.06And then once the chromosome is chosen,
00:26:11.09there's a very intriguing event
00:26:14.06in which silencing nucleates onto the X inactivation center
00:26:17.28and then spreads outwardly from that chromosome,
00:26:21.18to cover the entire 160 megabase or so chromosome.
00:26:26.19So, all of this is over by embryonic day six and a half, let's say,
00:26:31.29which is shortly after implantation in the mouse.
00:26:34.22Now, thereafter, the chromosome goes into a maintenance phase,
00:26:38.18a very important maintenance phase,
00:26:40.25in which the same X chromosome
00:26:44.09remains the inactive one for the lifetime of that female.
00:26:49.10So, once it's chosen,
00:26:51.05forevermore that chromosome will be inactive.
00:26:54.03Alright.
00:26:55.29So, many of us in the field have been
00:26:58.08trying to understand the mechanisms by which these steps take place.
00:27:02.25So, I'm gonna just outline some of the conceptual challenges
00:27:05.16associated with studying these problems.
00:27:09.00We'll begin with the counting problem.
00:27:11.14So, X chromo...
00:27:13.19X chromosome counting takes place in the blastocyst,
00:27:18.10shortly after the paternal X chromosome reactivates.
00:27:22.06Okay, so remember back to the life cycle of the X chromosome.
00:27:24.26So, the paternal X chromosome reactivates.
00:27:27.17And then a counting mechanism comes into play,
00:27:30.23in which every cell in the epiblast
00:27:35.11makes a determination of the sex chromosome number.
00:27:39.20So, this takes place around the time
00:27:41.28that there are about 20 cells in the epiblast,
00:27:45.19so we say it's a cell-autonomous decision.
00:27:47.25So, how does this all work?
00:27:49.17Now, it turns out that
00:27:52.17cells aren't actually counting the absolute number of X chromosomes,
00:27:55.23because that wouldn't make any sense.
00:27:57.17Cells are actually measuring
00:28:00.22the number of X chromosomes
00:28:03.02relative to the total genome content,
00:28:05.01so, the number of autosomes.
00:28:06.29So, we call that the X-to-autosome ratio.
00:28:09.23And it's been empirically observed
00:28:13.11by pioneers in the field
00:28:15.03that every cell
00:28:17.24-- if it's diploid, okay? --
00:28:20.02is allowed to keep one X chromosome active.
00:28:22.07So, cells follow this so-called n-1 rule,
00:28:26.07in which all X chromosomes are inactivated
00:28:29.03except for that one privileged, active X.
00:28:33.14Alright.
00:28:35.08So then, let's see what happens in males.
00:28:37.23So, males have an X-to-autosome ratio
00:28:40.10of 0.5.
00:28:42.06So, if the male cell is diploid,
00:28:45.01it keeps its one active X chromosome.
00:28:48.22If the male is tetraploid
00:28:50.24-- so it has twice the genome content,
00:28:52.22so we call it 4n, here --
00:28:54.22then it's allowed to keep two active X's.
00:28:58.05There are no supernumerary X chromosomes,
00:29:00.20so males, whether it's diploid or tetraploid,
00:29:02.19will not undergo X inactivation.
00:29:05.16So then, contrast that with what happens in females,
00:29:08.17where the X-to-autosome ratio is 1.0.
00:29:12.07So, if the female is diploid and following the n-1 rule,
00:29:15.13it will inactivate one of its two X chromosomes.
00:29:19.15Now, if she has twice the genomic content,
00:29:23.27then she will keep two X chromosomes active,
00:29:26.28and silence the two additional X chromosomes
00:29:31.02in her cells,
00:29:33.09Alright.
00:29:35.05So, now let's go back to the diploid,
00:29:37.06and assume for a moment that this female is a mosaic,
00:29:39.06and she has a mixture of cells that are dip...
00:29:41.14that are... that have two X chromosomes,
00:29:43.18and some that have three X chromosomes.
00:29:47.15Then, by the n-1 rule,
00:29:49.25she would inactivate two,
00:29:52.08the two blue chromosomes,
00:29:54.02out of her three.
00:29:55.26And suppose, again,
00:29:58.12that she actually had a mixture of cells
00:30:00.10that had four X chromosomes.
00:30:02.17And in this situation, by the n-1 rule,
00:30:05.02she would inactivate three of her four X chromosomes.
00:30:09.05So, very complicated.
00:30:10.27Alright.
00:30:12.22Now, how do we envision all of this happening
00:30:15.16at the molecular level.
00:30:16.27That's the big question.
00:30:18.26So, we believe that counting
00:30:21.17is really a titration of
00:30:24.09X-linked and autosomal factors.
00:30:27.06Okay?
00:30:28.13So, those go to make up the X-to-autosome ratio.
00:30:32.03And so the question is, what are these numerators?
00:30:34.27What are these X-linked factors?
00:30:36.18And what are the denominators,
00:30:38.09the autosomal factors?
00:30:40.14So, here's a model that's very popular in the field,
00:30:44.29and the idea is as follows,
00:30:48.17where in male cells, the X chromosome
00:30:51.09will make a limited set...
00:30:53.18or it will make a set of these so-called numerators,
00:30:55.27which are in green,
00:30:58.05in limited quantities.
00:30:59.22Okay?
00:31:01.07And likewise, the autosomes,
00:31:02.28shown in pink here,
00:31:04.16will make these denominators.
00:31:06.13They're also produced in limited quantities.
00:31:09.11And so the numerators and the denominators
00:31:12.09-- so, the green and red factors --
00:31:14.05will titrate each other out to form
00:31:17.06what the field calls a blocking factor,
00:31:19.15which will then go and bind to one
00:31:23.14-- the one and only --
00:31:25.10inactivation center in the cell,
00:31:27.00and prevent that inactivation center
00:31:30.01from firing to induce X chromosome inactivation.
00:31:32.12So then, the male is spared of X chromosome silencing.
00:31:37.10So, then a similar thing would happen in the female.
00:31:40.28So, she also makes red and green factors,
00:31:43.28and the red and green factors titrate each other
00:31:46.11to form this blocking factor.
00:31:47.27And then the blocking factor
00:31:49.26goes and sits on one inactivation center,
00:31:53.03prevents a firing in that center,
00:31:54.15and that chromosome remains active.
00:31:56.26Now, again,
00:31:58.26the early pioneers in the field
00:32:01.10felt that the blocking factor was sufficient to explain
00:32:04.08X chromosome counting,
00:32:05.23because then all of the remaining X chromosomes
00:32:08.21that weren't lucky enough to get the blocking factor
00:32:10.21would undergo silencing by default.
00:32:14.11Okay.
00:32:15.23So, now, that model certainly works very well,
00:32:17.18and it could in fact be what's happening in cells.
00:32:20.02But we'd like to think that biology is purposeful, right?,
00:32:24.25so that... nothing happens in life sort of randomly,
00:32:30.14or by default.
00:32:32.00And that there should be
00:32:34.02a purposeful inactivation of the remaining X chromosomes.
00:32:38.00And so, in fact,
00:32:39.24there is something different here about the female.
00:32:42.04So, has one extra X chromosome,
00:32:44.17which means that she's producing extra green factors.
00:32:48.23Now, these green factors go untitrated by the blocking factor.
00:32:54.15And so we suggest that these
00:32:57.10additional green factors
00:32:59.11would go and make their own complex
00:33:02.02-- a complex that we call the competence factor.
00:33:05.01And that factor would then go
00:33:07.25and bind to the remaining X chromosome,
00:33:10.12and purposefully induce that inactivation center
00:33:15.25to fire and initiate the cascade of silencing.
00:33:20.24Okay.
00:33:22.05So, we call that the two factors hypothesis.
00:33:25.09And I'm gonna have a lot more to say about that
00:33:27.08in the second lecture.
00:33:28.29But for now, I'd like to move on to the next conceptual challenge,
00:33:31.18which is allelic choice.
00:33:33.27And this is one of my favorite problems.
00:33:35.27Because what this allelic choice really means
00:33:41.07is that that decision has to take place instantaneously.
00:33:45.17In order for it to be a robust mechanism,
00:33:47.07it has to be mutually exclusive,
00:33:50.11and irreversible.
00:33:51.25Right?
00:33:52.24Because we know that once a decision is made,
00:33:54.25well, you don't go back on that decision.
00:33:56.21And the same X chromosome remains inactive
00:33:59.22for the lifetime of that female.
00:34:02.12Okay, so here's the big problem.
00:34:04.08How do we make the right choice?
00:34:06.09And how do we make that choice in a mutually exclusive fashion?
00:34:09.26So, we postulate that
00:34:15.12there has to be a mechanism of communication
00:34:18.16between the two X chromosomes
00:34:21.01-- so, trans, we say... trans communication --
00:34:24.24such that when one chromosome is chosen as the inactive one,
00:34:30.04this other X chromosome has to
00:34:33.13instantaneously become the active X chromosome.
00:34:37.15Okay.
00:34:38.27So, the idea is that one hand
00:34:41.19has to know what the other hand is doing
00:34:43.18at all times.
00:34:44.27And in fact, we have that problem with our brain.
00:34:46.26Because, you know, we have two halves to our brain,
00:34:49.02and one... the right brain controls the left hand,
00:34:51.03and the left brain controls the right hand.
00:34:54.06And so how do we communicate between the two halves of the brain?
00:34:57.07And so the brain does it very well,
00:34:59.03through a bridge called the corpus callosum.
00:35:01.27And so we postulate that there has to be
00:35:06.13a bridge between the two X chromosomes as well,
00:35:08.18that allows the left hand
00:35:11.10to know what the right hand is doing.
00:35:13.12Okay?
00:35:14.26So, in the next lecture,
00:35:16.06I'm gonna be talking a lot about this hypothesis
00:35:17.27that that communication bridge
00:35:20.02is built during a very transient pairing event
00:35:24.29between the two X chromosomes.
00:35:27.20And so, briefly,
00:35:29.19the idea here is that prior to X inactivation
00:35:31.16we observe empirically that the two X chromosomes
00:35:34.11sort of do their own dance.
00:35:37.07Then, at the onset of... just before...
00:35:39.29actually, just before the initiation of X inactivation,
00:35:43.12we see that the two X chromosomes
00:35:46.19come together in three-dimensional space,
00:35:48.23and they briefly touch each other
00:35:50.25at the X inactivation center.
00:35:52.21And then, when they come apart again,
00:35:54.20one X chromosome is active
00:35:57.20and the other one has become inactive.
00:36:00.14Okay, so we have suggested
00:36:04.24that this mechanism of pairing
00:36:06.21may underlie the ability of cells
00:36:09.00to determine choice instantaneously,
00:36:11.00in a mutually exclusive fashion,
00:36:12.24and in an irreversible fashion.
00:36:15.28So, more about that in Lecture 2.
00:36:18.24Alright.
00:36:20.05So then, the final conceptual challenge
00:36:22.01is how the chromosome
00:36:24.06initiates and spreads silencing.
00:36:26.02So, as we briefly touched upon already...
00:36:29.21so, inactivation starts within the XIC,
00:36:32.21the X inactivation center,
00:36:34.15and then it spreads throughout the rest of that chromosome
00:36:40.22without touching the other X chromosome,
00:36:43.07or for that matter any other chromosome in the cell.
00:36:47.03Now, that's a big conceptual problem,
00:36:48.18because we know that proteins,
00:36:50.13and just about all biological factors that we know,
00:36:53.18can diffuse in some way.
00:36:56.00And so... and the two X chromosomes
00:36:58.18and all the other chromosomes
00:37:00.08lie in the same nucleoplasm.
00:37:01.20And so, how is it that we
00:37:04.06render all other chromosomes immune to the inactivation process
00:37:06.23except for this one?
00:37:08.19Alright.
00:37:09.23So, that's the conceptual challenge.
00:37:11.15And so this gene, Xist,
00:37:14.07discovered many, many years ago now,
00:37:16.19is a 17-20 kilobase long non-coding RNA.
00:37:21.02And it has a distinction of being
00:37:23.18the only gene which is transcribed
00:37:26.13from the inactive X chromosome.
00:37:27.29Now, all other genes are transcribed from the active X chromosome,
00:37:30.07by definition.
00:37:31.19This gene is made only from the inactive one.
00:37:36.10And from experiments done many, many years ago,
00:37:38.24we know that Xist is absolutely essential
00:37:42.15for X chromosome inactivation.
00:37:44.09And this image, here,
00:37:46.06is an RNA fluorescence in situ hybridization
00:37:48.05that shows that Xist,
00:37:49.23which is here in red,
00:37:53.08"coats", in quotes,
00:37:55.01the inactive X chromosome
00:37:57.14without diffusing to any of the other chromosomes
00:37:59.24in the same cell,
00:38:01.08including the other active...
00:38:03.00the other X chromosome.
00:38:04.23So, we say that the RNA localizes strictly "in cis".
00:38:09.02Okay.
00:38:10.21Now, we know that this property
00:38:14.03originates from the X inactivation center.
00:38:17.05So, we did these transgenesis experiments
00:38:20.20I showed you a few minutes ago,
00:38:23.00in which we move the Xist gene onto an autosome,
00:38:26.02and now that autosome gets coated by Xist RNA.
00:38:30.24And again,
00:38:32.10none of the other chromosomes get coated by Xist RNA.
00:38:34.09So, the property is derived
00:38:37.28from the X inactivation center,
00:38:39.16and there do not appear to be other X-specific elements
00:38:42.20responsible for this behavior.
00:38:46.20Alright.
00:38:48.03So, Xist starts... here at the top, Xist is synthesized.
00:38:52.01And it spreads in three dimensions
00:38:55.01to the rest of the chromosome.
00:38:57.03And then it does at least three things that we're aware of.
00:39:01.05So, firstly, it has to
00:39:03.08push away all of the existing activating factors.
00:39:06.06You know, you're trying to convert an active chromosome
00:39:08.21to an inactive chromosome,
00:39:10.03so one of the first things that Xist will do
00:39:12.05is actually push away the activating factors.
00:39:13.23And then it recruits silencing factors
00:39:18.10to that chromosome,
00:39:20.01to begin the process.
00:39:21.18And at the same time,
00:39:23.25it is changing the topology of that chromosome,
00:39:26.12changing the three-dimensional architecture of that chromosome.
00:39:29.11And the net result of all of this
00:39:33.22is a highly stable, a very robust,
00:39:36.13gene silencing mechanism.
00:39:38.10So, we're gonna say a lot more about that in Lecture 3.
00:39:43.25Now, before I conclude this first lecture,
00:39:46.05I would just like to say a few words
00:39:48.21about some efforts that are underway at the present time
00:39:52.15to leverage our understanding of how X inactivation works
00:39:57.20to treat human disease.
00:39:59.21So, this is a very active area of research
00:40:02.13because, as I mentioned before,
00:40:05.10there are more than 200 disease genes
00:40:08.03on the X chromosome.
00:40:09.23And here, though,
00:40:12.09through the lifetime of the female,
00:40:14.13the inactive X chromosome lies completely dormant.
00:40:17.09So, the idea is...
00:40:20.20well, here's a silent chromosome,
00:40:22.05but it's a silent chromosome that carries
00:40:24.091000 very good genes,
00:40:26.04the genes that could function.
00:40:28.02So, it's a reservoir of genes that could potentially
00:40:30.06be reawakened for therapeutic purposes,
00:40:32.14especially when there is a mutation expressed
00:40:36.25on the active X chromosome.
00:40:38.11So, what we'd like to know is
00:40:40.07whether we could unlock the inactive X chromosome
00:40:43.04to treat X-linked disease,
00:40:45.00using all of the knowledge that we have gathered
00:40:47.29over the past half-century.
00:40:50.27Okay.
00:40:52.07And the poster child for this X-reactivation platform
00:40:54.12is Rett syndrome.
00:40:56.20So, Rett syndrome is a devastating disorder
00:40:59.09that's caused by a mutation in the protein called MECP2,
00:41:03.03which is carried on the X chromosome.
00:41:05.04And this disorder arises in girls
00:41:07.21who unfortunately inherit one mutant copy of MECP2.
00:41:14.01So, she's a mosaic.
00:41:15.26She's... 50% of her cells, approximately,
00:41:18.12will express normal MECP2,
00:41:20.15and the other half of cells
00:41:23.02will not have a functioning MECP2.
00:41:26.12So, the girls are born normal,
00:41:28.06but then during the first year of life,
00:41:31.07they begin to manifest very severe symptoms.
00:41:34.20They lose their learned abilities.
00:41:36.14They develop severe intractable seizures,
00:41:38.25get repeated lung infections,
00:41:40.24and they have autism and repetitive behaviors.
00:41:43.26And many of the girls never learn to walk or talk.
00:41:48.22So, it is a very devastating, life-long disorder.
00:41:52.18And there are currently no disease-specific treatments.
00:41:56.06Okay.
00:41:57.24But there was a very inspiring study
00:41:59.27from Adrian Bird's lab about ten years ago,
00:42:01.29in which they showed that, at least in a mouse model,
00:42:07.07they could reverse neurological defects of Rett syndrome
00:42:10.07if they could resupply,
00:42:13.07or give back to the brain,
00:42:15.02normal quantities of MECP2,
00:42:18.15even after the onset of disease symptoms.
00:42:22.01And so that's led many of us to ask this question
00:42:25.08-- can we use our knowledge of X-inactivation
00:42:28.06to reactivate the inactive X chromosome,
00:42:31.01and thereby restore expression of MECP2 protein?
00:42:34.16And maybe partially treat Rett syndrome?
00:42:38.18So, we've demonstrated that
00:42:40.175-10% of normal MECP2 expression in the brain
00:42:44.09can have significant phenotypic impact.
00:42:48.03So, for example, here in this mouse model,
00:42:51.08a mouse that has no MECP2 function at all
00:42:54.22will live about 76 days.
00:42:57.05But if we give it even as little as 1% MECP2,
00:43:02.06those mice will live a month longer.
00:43:05.13Now, if we were to give them 5% MECP2 expression
00:43:09.18in the brain,
00:43:11.10those mice will have their lives extended
00:43:13.28by 2-3-fold.
00:43:15.10And furthermore,
00:43:17.23if we went up to 10-20% MECP...
00:43:20.20MECP2 expression in the brain,
00:43:23.06their life span is extended up to 5-fold,
00:43:26.28with a proportionate improvement to their neuromotor function.
00:43:31.04And so, the point is that
00:43:33.12a small amount of MECP2 protein
00:43:35.03can go a long ways towards good brain function.
00:43:40.06And so with that in mind,
00:43:41.26what many of us are trying to do in the field, now,
00:43:44.12is to partially reactivate the X chromosome,
00:43:47.01and thereby boost expression of MECP2,
00:43:49.25specifically in the brain.
00:43:52.14What we have identified in...
00:43:55.12within the last year or so
00:43:57.09is a specific therapeutic co*cktail
00:44:01.12that would boost MECP2 significantly inside cells.
00:44:06.15And so this is a combination
00:44:08.20of an anti-Xist molecule
00:44:11.09and an anti-DNA methylation molecule.
00:44:14.29And what you can see in these graphs, here,
00:44:17.07is that the anti-Xist molecule by itself
00:44:19.24does essentially nothing.
00:44:23.00And the anti-methylation molecule by itself
00:44:26.29does ever so little.
00:44:29.08But if we combine the two of them, we get a huge boost,
00:44:32.25a 30,000-fold boost,
00:44:35.02in MECP2 expression.
00:44:37.22And so we're presently testing this drug candidate
00:44:41.02in preclinical models
00:44:43.00in hopes of identifying, eventually,
00:44:45.14a clinical candidate that could be
00:44:49.01translated to the clinic.
00:44:51.06Okay.
00:44:52.21So, that then brings us to the end of Lecture 1.
00:44:57.18And what I've told you in this lecture
00:44:59.16is that X inactivation is an essential developmental process.
00:45:03.07It in many ways exemplifies the challenges
00:45:07.00that we as scientists have
00:45:09.29in trying to understand epigenetic regulation.
00:45:13.16And I mentioned also that
00:45:16.08our knowledge of X inactivation
00:45:18.22may eventually be leveraged
00:45:20.25to treat certain X-linked diseases.

00:00:15.06Welcome back, everyone.
00:00:16.22So, again, I'm Jeanie Lee.
00:00:18.18I'm a Professor of Genetics at Harvard Medical School,
00:00:21.21and I'm also a faculty member in the Department of Molecular Biology
00:00:24.17at Massachusetts General Hospital.
00:00:27.00Now, in Lecture 1,
00:00:28.22I gave an overview of X chromosome inactivation.
00:00:31.23And what we're gonna do now in Lecture 2
00:00:33.27is a deeper dive into the initiation phase of inactivation,
00:00:38.27namely how cells count
00:00:41.15and then make the correct choice
00:00:43.29of active and inactive chromosomes.
00:00:46.22So, here again are the different steps of X inactivation.
00:00:49.20And by way of review, there's a counting mechanism followed by a choice mechanism,
00:00:53.27and then the initiation of silencing.
00:00:57.28So, we're gonna focus first on the counting step.
00:01:01.25So, again, this takes place in the blastocyst,
00:01:04.25shortly after the paternal X chromosome is reactivated.
00:01:08.07And every cell makes its own determination
00:01:11.14of the X chromosome number.
00:01:13.14So, it's taking place around the time
00:01:15.15that the epiblast has 20 cells or so.
00:01:19.18Alright.
00:01:20.29And we discussed how it's really the X-to-autosome ratio
00:01:24.09that cells are sensing,
00:01:25.23rather than the absolute number of X chromosomes.
00:01:27.27And we also mentioned that cells
00:01:31.05follow the n-1 rule per diploid content,
00:01:34.00so that males who have an X-to-autosome ratio of 0.5
00:01:39.02will not inactivate any chromosomes,
00:01:42.04regardless of whether it's diploid or tetraploid,
00:01:45.22with twice the genomic content.
00:01:47.26Whereas the female,
00:01:49.28with an X-to-autosome ratio of 1.0,
00:01:52.10will inactivate one of her two X chromosomes,
00:01:55.24and if she's tetraploid she'll inactivate
00:01:58.11two out of those four X chromosomes.
00:02:00.23And furthermore, if the diploid had three X chromosomes,
00:02:04.10it will inactivate two out of the three.
00:02:07.10And if she has four X chromosomes,
00:02:10.21she'll inactivate three out four.
00:02:12.27And so the point is that cells follow this n-1 rule
00:02:16.06in a diploid content,
00:02:17.26and every cell makes the determination for itself.
00:02:21.07And we also mentioned that
00:02:24.20counting is most likely a titration of X-linked
00:02:27.08and autosomal factors.
00:02:28.25We mentioned that the numerators
00:02:30.21are produced from the X chromosome,
00:02:33.22in the form of this green blob.
00:02:35.26And the autosomes also produce their own factors
00:02:39.17-- we call them denominators.
00:02:41.12And then the red and green factors
00:02:43.15will titrate each other out,
00:02:45.04and form a blocking factor
00:02:47.03that then sits on one X chromosome,
00:02:50.08at the inactivation center,
00:02:51.20and prevents that inactivation center
00:02:54.19from initiating the cascade of inactivation.
00:02:58.03So, we see that in the male cell.
00:03:00.15And we also see that in the female cell,
00:03:02.20where the same factors titrate each other out
00:03:06.11to form a hypothetical blocking factor,
00:03:08.26which then sits on the...
00:03:11.25one X chromosome, preventing that inactivation center from firing,
00:03:15.11giving rise to this privileged, active X chromosome.
00:03:19.27And we mentioned that one hypothesis
00:03:22.21is that the remaining X chromosomes
00:03:25.11would then undergo an inactivation by default.
00:03:28.22So, that's certainly one viable viewpoint.
00:03:31.10However, we favor the idea that there is a purposeful inactivation
00:03:35.17-- not something that happens by default.
00:03:37.17Because, in fact, the female produces
00:03:40.29an extra copy of these green factors,
00:03:44.12since she has one extra X chromosome.
00:03:46.21And that green factor is not titrated by the blocking factor,
00:03:50.18so we propose that that factor
00:03:53.25goes and forms this additional complex
00:03:58.01called a competence factor,
00:04:00.08which has to sit on the remaining X chromosome
00:04:04.00to purposely induce the initiation of inactivation.
00:04:08.22So, that's the two factors hypothesis.
00:04:11.23Okay.
00:04:13.12So, then we get down to,
00:04:15.14what are these molecular factors
00:04:18.29that make up the X, right?
00:04:21.02And what are the factors that make up
00:04:22.29the A of the X-to-autosome ratio?
00:04:25.09And here we'll start with the numerator, the X.
00:04:27.29Alright.
00:04:29.04So, in principle, without knowing all that much
00:04:31.16about what these factors are,
00:04:33.08we can say that the factor
00:04:35.22has to be produced from the X inactivation center
00:04:37.25-- from those transgenesis experiments that I showed you before.
00:04:40.21And we believe that that factor
00:04:43.20has to escape X inactivation.
00:04:46.02So, I haven't mentioned before,
00:04:47.26but a number of genes on this chromosome
00:04:51.04are actually immune to the influence of Xist.
00:04:55.26They escape silencing.
00:04:57.17And we believe that numerators
00:04:59.22have to escape X inactivation
00:05:02.01in order to serve as a dosage-sensitive readout
00:05:05.29of that chromosome.
00:05:08.17And furthermore, that factor has to be diffusible.
00:05:11.16So, in order to titrate away autosomal factors,
00:05:13.17it has to be able to move around in the nucleus.
00:05:16.24And then finally,
00:05:18.18it has to act at the X inactivation center,
00:05:20.25which is where the Xist gene ultimately resides.
00:05:24.23And then, most importantly,
00:05:26.20the math has to work.
00:05:28.10And what I mean by that is,
00:05:30.14if something were truly a numerator,
00:05:33.13then when we take away one copy of that X-linked numerator,
00:05:38.16a female's cells should start to behave like male cells
00:05:42.22and block X inactivation.
00:05:44.22Right?
00:05:46.09So, she should think that she's a male cell,
00:05:48.01because she's missing the extra X-linked factor.
00:05:50.27And conversely,
00:05:53.14if we were to give a male cell extra copies of a numerator,
00:05:57.05that male cell has to start to behave like a female cell
00:06:00.14and start undergoing X chromosome inactivation.
00:06:04.06Okay.
00:06:05.19So, these are the rules.
00:06:07.08Now, a while back,
00:06:09.10we started to suspect this non-coding gene, Jpx,
00:06:12.21which lies just on the other side of Xist.
00:06:15.23It's an X-linked Xic product.
00:06:18.16We know that Jpx levels
00:06:21.27increase about tenfold during the process of X inactivation.
00:06:24.29And it occurs in the same timeframe that Xist
00:06:29.01is getting upregulated on that chromosome,
00:06:31.15and we know that it escapes X inactivation.
00:06:33.29It's one of the few genes that will escape inactivation.
00:06:36.27And from the transgenesis experiments...
00:06:39.26I told you before, when we move this region
00:06:41.27and put it on an autosome,
00:06:43.14that autosome behaves like an X chromosome
00:06:45.18and undergoes inactivation.
00:06:47.09However, if we were to make a transgene
00:06:50.04that's missing this Jpx,
00:06:53.14that transgene could no longer inactivate the autosome.
00:06:58.13Okay.
00:06:59.29So, we put this to the test.
00:07:03.16So, here... this is an RNA fluorescence in situ experiment,
00:07:06.23in which we're looking at expression of Xist,
00:07:09.06which is shown here as a pink dot,
00:07:12.00and it's coating the X chromosome.
00:07:14.16So, in wildtype cells we see a very robust expression of Xist RNA.
00:07:19.15Now, if we, in the female cell,
00:07:21.29deleted just one copy of Jpx...
00:07:25.07so, there are two copies normally...
00:07:27.06we just take away one copy of Jpx...
00:07:29.00you see that these cells no longer produce that
00:07:32.14large, robust cloud of RNA
00:07:35.20that coats the inactive X chromosome.
00:07:38.05However, if we then take a copy of Jpx
00:07:41.16and insert it into another chromosome, an autosome,
00:07:46.03we rescue this expression of Xist
00:07:50.15in the same female cells.
00:07:52.15Okay.
00:07:53.22So, these experiments tell us two very important things.
00:07:56.13First of all, Jpx is a dosage-sensitive element.
00:08:01.18So, by removing one copy of Jpx
00:08:04.25in these female cells,
00:08:06.17the female cells start to behave like a male cell.
00:08:09.26And furthermore, Jpx is diffusible,
00:08:13.19because we put the gene on an autosome.
00:08:16.02That autosomally produced Jpx will rescue Xist expression
00:08:20.18on the X chromosome.
00:08:22.17And then we did the converse experiment,
00:08:24.27where, in male cells now,
00:08:27.06we insert extra copies of Jpx.
00:08:30.04Male cells normally don't produce Xist at all,
00:08:32.13so you don't see a big Xist spot in green.
00:08:35.04But when we insert extra copies of Jpx
00:08:38.22into these male cells,
00:08:40.16you start to see Xist expression go up, okay?
00:08:45.03So, that suggests that Jpx may in fact
00:08:48.00be a candidate for a numerator factor.
00:08:51.15And in this experiment, here,
00:08:53.29we're demonstrating that Jpx is acting as a diffusible RNA,
00:08:58.07one of these long non-coding RNAs.
00:09:00.06And it's not simply the genetic element, the DNA,
00:09:04.04which is responsible for this counting act.
00:09:06.18And we know that because
00:09:09.26when we introduce these things called shRNAs...
00:09:12.11this is a technology that allows us to degrade the RNA
00:09:15.17when we introduce the shRNA into cells,
00:09:18.07without actually touching the underlying gene.
00:09:22.00Okay, so when we introduce these RNA degradation factors,
00:09:25.15we see that Xist could no longer be upregulated,
00:09:29.25like in the wildtype female cell,
00:09:32.06or the untouched female cell.
00:09:34.15So, this experiment tells us that
00:09:38.02Jpx is a diffusible element, acting as a non-coding RNA.
00:09:43.05Okay.
00:09:44.24So, the idea then is that
00:09:47.09Jpx is one of these green factors
00:09:49.26that's being produced by the X chromosome,
00:09:51.25and that it is titrating away the autosomal blocking factor.
00:09:56.09And then the untitrated Jpx factors
00:09:59.22would be the one that sits on the remaining inactive X
00:10:04.19to induce the firing of that inactivation center.
00:10:07.21Okay.
00:10:09.00So then, we turn our attention to,
00:10:10.29what are the pink factors?
00:10:12.15What are these autosomal factors
00:10:14.15which are getting expressed to titrate Jpx?
00:10:18.19Now, here, we began to suspect
00:10:21.03a protein called CTCF.
00:10:23.19Now, CTCF is a very famous protein
00:10:26.01because it does a lot of different things.
00:10:28.02It has been shown to be a critical chromosome architectural factor.
00:10:32.15It was first identified by Victor Lobanenkov
00:10:34.23as an 11-zinc finger transcription factor
00:10:38.01that can take two distant genetic elements
00:10:41.26and bring them together to form a loop,
00:10:44.08as shown here,
00:10:45.23and can regulate enhancer-promoter interactions.
00:10:50.03And more recently, CTCF has been shown to
00:10:53.20reside at the border of these chromosomal topological structures
00:10:56.17called TADs,
00:10:58.00and I'll say a lot more about that in lecture number 3.
00:11:00.22Okay.
00:11:03.07But importantly, we've known for quite some time that
00:11:05.25CTCF occupies discrete positions
00:11:08.29at the X inactivation center
00:11:11.11and plays an important role in a number of different processes.
00:11:15.16So, for example, here
00:11:18.29CTCF binds to the Xist promoter at a number of positions
00:11:21.23that are shown here in red.
00:11:23.25And we know that at these sites
00:11:26.07CTCF is serving as a repressor of the Xist gene.
00:11:31.01And then, as cells go through X inactivation,
00:11:33.12these binding sites here
00:11:35.16-- shown, again, in red --
00:11:37.14pretty much stay the same.
00:11:39.20They remain bound.
00:11:41.17Except for one... at one location,
00:11:43.23the so-called P2 location.
00:11:45.29Now, at this position,
00:11:48.17CTCF binding actually goes down during X inactivation.
00:11:53.25So, that was really interesting.
00:11:55.17And we wanted to know,
00:11:57.07because there are two X chromosomes,
00:11:58.24from which X chromosome CTCF was getting removed.
00:12:01.22So, for that, we had to perform
00:12:04.02an allele-specific analysis.
00:12:05.29You know... so, that's an analysis that allows us to
00:12:08.10tell the difference between the future active
00:12:10.11and the future inactive chromosomes.
00:12:12.17And the long and the short of this is that it is from the future inactive...
00:12:17.19the chromosome which will become inactivated...
00:12:19.22that's where CTCF is losing its binding.
00:12:24.08Okay. So, during X inactivation,
00:12:28.08CTCF at P2 is retained only on the active X chromosome.
00:12:31.26And we know that its role is to
00:12:34.24block the expression of this critical silencing factor called Xist.
00:12:39.14So, what I've told you, then,
00:12:41.13is that CTCF is an autosomal factor.
00:12:45.14It represses Xist expression.
00:12:48.00And at the same time,
00:12:49.25I've told you that this non-coding RNA that's X-linked
00:12:54.12induces Xist expression.
00:12:56.26And so, with one being autosomal,
00:12:58.25the other one being X-linked,
00:13:00.16and doing opposite things,
00:13:02.19we wondered whether these two factors
00:13:05.13could be functionally interacting with each other,
00:13:07.16and be part of that titration mechanism
00:13:10.04that I referred to earlier,
00:13:12.07part of the X inactiv... the X-to-autosome ratio.
00:13:16.03So, indeed, we learned that CTCF is an RNA-binding protein.
00:13:22.01That was not previously known to bind RNA.
00:13:24.25But in this context,
00:13:26.25it is a very good RNA-binding protein.
00:13:28.20In fact, it prefers to bind RNA over DNA.
00:13:31.22So, you can see from the same sort of gel shift analysis, here...
00:13:35.03except that this time we're using Jpx RNA,
00:13:38.15and you see that very robust shift,
00:13:40.10indicating a high-affinity binding
00:13:43.04between the RNA and CTCF protein.
00:13:45.20And so, in fact, we can biochemically
00:13:48.26measure the affinity of this complex
00:13:51.05by measuring the dissociation constant.
00:13:54.16And that Kd is less than one nanomolar.
00:13:57.14So, CTCF is a very good RNA binding protein,
00:14:00.20much better than binding to...
00:14:03.15its binding to DNA,
00:14:05.26where the dissociation constant is more than 20 nanomolar.
00:14:12.23Okay.
00:14:14.21So, then we have this idea that
00:14:17.13CTCF may be getting competed away from the promoter
00:14:20.10by this non-coding RNA, Jpx,
00:14:24.01and that may underlie this titration mechanism.
00:14:27.06And so to test that,
00:14:29.10we mixed together purified components of P2 DNA,
00:14:34.00CTCF bound to the P2 DNA,
00:14:37.12and increasing concentrations of this Jpx RNA.
00:14:41.08And what we see here in this gel shift analysis
00:14:44.28is that CTCF gets pulled away from the DNA
00:14:49.10by Jpx RNA.
00:14:52.09Okay.
00:14:53.22So, we can do the same sort of competition experiment
00:14:56.17inside of cells.
00:14:58.12Now, what I've shown you so far
00:15:00.23occurred within a test tube, right?,
00:15:02.17but we can do this sort of thing inside a cell as well.
00:15:05.23So, here, we're overexpressing CTCF
00:15:09.10-- that's the repressor of Xist --
00:15:11.03and you can see that when we do that
00:15:13.20the cells no longer upregulate this green cloud of Xist.
00:15:17.08So, here's wildtype, as you can see.
00:15:19.20But in the overexpression system,
00:15:22.12we no longer see Xist clouds.
00:15:26.00However, we can overcome these extra quantities,
00:15:31.08if you will,
00:15:33.14of CTCF by giving the cells extra Jpx RNA.
00:15:37.23And so that's what we've done here.
00:15:39.05And you see that these green spots come back.
00:15:42.07Okay.
00:15:44.01So, that supports this idea that
00:15:47.22CTCF and Jpx RNA are functionally interacting with each other
00:15:51.16and titrating each other inside of cells.
00:15:55.22So, what we propose, then,
00:15:58.26is a functional antagonism between CTCF and Jpx RNA.
00:16:03.04So, prior to X inactivation,
00:16:05.04CTCF sits very robustly at the 5' end of Xist,
00:16:08.24where it blocks the expression of Xist.
00:16:11.24And then, at the onset of X inactivation,
00:16:15.09what we have empirically measured is that Jpx RNA
00:16:19.28increases in expression by tenfold.
00:16:22.10And when it crosses a certain threshold,
00:16:25.05as it will do only in female cells
00:16:27.03-- because we have twice the number of Jpx copies as male cells --
00:16:32.07the Jpx RNA binds to CTCF
00:16:36.03and titrates it away from one Xist promoter.
00:16:40.21And that act enables Xist RNA
00:16:44.14to be upregulated on that same chromosome.
00:16:47.24Okay.
00:16:49.06So, that's how we're presently thinking about
00:16:52.00this functional antagonism
00:16:54.08and about the X-to-autosome ratio.
00:16:57.04What I'd now like to turn your attention to
00:16:59.23is the second step of X inactivation,
00:17:02.00which is allelic choice.
00:17:04.05And I mentioned in the first lecture
00:17:06.18that this is a conceptually very challenging problem,
00:17:09.01because, here, choice has to be random.
00:17:14.10It has to be instantaneous,
00:17:17.17mutually exclusive,
00:17:19.04and completely irreversible.
00:17:21.03Okay.
00:17:23.03So, how do we make the right choice.
00:17:24.22And again, we believe that there is a communication
00:17:27.19between the two chromosomes,
00:17:29.01such that when one chromosome is chosen as the inactive one
00:17:31.19the other one is instantaneously the active chromosome.
00:17:36.28So, this mutually exclusive choice
00:17:39.14-- which is what we call it --
00:17:40.28requires two genetic loci at the X inactivation center.
00:17:44.25So, one is Xist's antisense repressor,
00:17:49.23called Tsix, shown here in yellow,
00:17:52.11and the other its enhancer,
00:17:54.21shown here in brown, called Xite.
00:17:57.23Okay.
00:17:59.12So, the region that's responsible for choice
00:18:01.09is this 15 kilobase domain
00:18:03.21that encompasses Tsix and Xite.
00:18:06.13And what we know
00:18:08.23-- going back to experiments done many, many years ago --
00:18:11.08is that prior to X inactivation,
00:18:13.22when the two chromosomes are active,
00:18:15.25the Tsix antisense RNA is expressed at very high levels,
00:18:20.22and its expression prevents Xist from turning up.
00:18:26.03Okay.
00:18:27.27But then, at the onset of X inactivation,
00:18:30.01what happens is that
00:18:32.27the antisense RNA disappears from one X chromosome.
00:18:36.07And when it disappears,
00:18:38.05Xist RNA is upregulated from that chromosome,
00:18:41.18leading to the formation of the inactive X.
00:18:44.23While on the other X chromosome,
00:18:47.07the action of the Xite enhancer, right?,
00:18:52.29allows Tsix to persist on that chromosome,
00:18:58.04so that the Xist gene continues to be repressed
00:19:01.12on that chromosome.
00:19:03.13And that chromosome stays active.
00:19:05.18So, the action of Tsix is essential for this mutually...
00:19:09.22for this allelic choice, with its persistence on the active...
00:19:15.09its persistence determining the active X chromosome,
00:19:19.24and its loss determining the inactive chromosome.
00:19:24.26Now, what we also demonstrated in these early studies
00:19:27.15is that we can genetically manipulate
00:19:30.15the choice decision
00:19:32.02by simply removing Tsix from one X chromosome.
00:19:35.00And when we do that,
00:19:36.21that chromosome is always the one
00:19:39.07that becomes inactivated.
00:19:41.01So, we can influence and manipulate which X chromosome
00:19:43.11will become the inactive one.
00:19:46.04Okay.
00:19:47.28So, then, what I told you is that
00:19:50.29normally cells can choose
00:19:53.20either one or the other X chromosome for inactivation.
00:19:56.21But very strangely,
00:19:59.01when we delete both copies of Tsix
00:20:01.27-- not just one, but both copies of Tsix --
00:20:04.14we see these additional cell types,
00:20:07.08where both X chromosomes are inactivated
00:20:10.13or neither X chromosome is inactivated.
00:20:14.16Right?
00:20:16.18So, it appeared to us here that the cells
00:20:20.29are undergoing some kind of a chaotic choice.
00:20:23.07Or maybe there's no choosing at all.
00:20:24.26You see all combinations as a result of losing this antisense repressor.
00:20:29.20So, this is a loss of mutually exclusive choice.
00:20:32.24And from that, we postulate that
00:20:35.26maybe there has been a loss of communication
00:20:38.26between the two X chromosomes,
00:20:40.20such that now cells...
00:20:42.21you know, really, the left brain doesn't know what the right brain is doing,
00:20:45.14going back to the analogy I drew in the first lecture.
00:20:49.19So, these experiments also tell us that
00:20:52.23the Tsix repressor is very important for that communication
00:20:56.08between the two chromosomes.
00:20:58.19Now, around the same time,
00:21:00.05we and the Heard lab made an interesting observation,
00:21:03.27which is that prior to X inactivation
00:21:08.24the X... two X chromosomes behave like they're not even aware of each other.
00:21:12.00But at the onset of X inactivation,
00:21:13.28one of the very first things that we see is that the chromosomes come together,
00:21:16.12and they briefly touch,
00:21:18.07just at the X inactivation center.
00:21:20.05And it's very brief.
00:21:22.03It happens probably in under 15 minutes,
00:21:24.03but let's say under 30 minutes.
00:21:25.23And then when they come apart again,
00:21:27.20one X chromosome is the active one;
00:21:29.16the other one is expressing Xist,
00:21:31.06so it's become the inactive one.
00:21:32.27It's almost as though the cells have flipped on a bistable switch
00:21:35.26as a result of pairing.
00:21:37.13And you can see this pairing event, here,
00:21:39.21by DNA fluorescence in situ hybridization,
00:21:42.10where you see two dots of the Xic
00:21:45.11coming close together in a certain timeframe
00:21:48.01during X inactivation.
00:21:49.17And so, because of this observation,
00:21:51.14we propose that the XX pairing process
00:21:55.06may serve as a bridge by which the two chromosomes
00:21:57.28can communicate with each other
00:22:00.19prior to the choice decision.
00:22:03.09Okay.
00:22:05.05So, in support of that idea, here...
00:22:08.14which we've shown...
00:22:10.27that the center responsible for pairing
00:22:14.04is the same region that's responsible for allelic choice.
00:22:18.21It's the same white bar that you show...
00:22:20.22that you saw a few slides earlier.
00:22:23.15So, this is a 15 kilobase region.
00:22:25.11And intriguingly,
00:22:27.25if we were to take this white line
00:22:30.03-- take this genetic region --
00:22:32.13and insert that into an autosome, far away,
00:22:35.23that autosome is now induced
00:22:39.04to pair with the X chromosome.
00:22:40.19So, this region, this very, very small region,
00:22:43.10is both necessary and sufficient
00:22:45.28to direct pairing.
00:22:48.08Alright.
00:22:49.24So, here are some real-life experiments.
00:22:51.21When we delete both copies of Tsix,
00:22:54.11the chromosomes no longer pair.
00:22:57.02And as I mentioned,
00:22:59.09there's a loss of mutually exclusive choice.
00:23:01.27On the other hand,
00:23:04.11if we insert extra copies of the pairing region into an autosome,
00:23:07.26which is shown here in blue,
00:23:09.28that autosome does something very strange,
00:23:12.19which is that it attracts one of the X chromosomes...
00:23:16.11one of the X chromosomes to come and pair with it,
00:23:19.01and in doing so it prevents the two X chromosomes
00:23:23.15from interacting with each other,
00:23:25.16and X inactivation is arrested.
00:23:28.29And so what these experiments tell us
00:23:31.17is that XX pairing is very important
00:23:33.21to somehow properly initiate X chromosome choice.
00:23:38.24So, we propose that pairing is a mechanism
00:23:41.20by which the two X chromosomes
00:23:44.09can break their epigenetic symmetry.
00:23:47.26So, prior to the onset of X inactivation,
00:23:50.26Tsix is expressed from both X chromosomes.
00:23:54.05And then the process of pairing results in the loss of antisense expression
00:23:59.28from one X chromosome,
00:24:01.19and it is from that chromosome that Xist becomes upregulated.
00:24:05.08And on the opposite X chromosome,
00:24:07.11Tsix persists,
00:24:09.19and that blocks the upregulation of Xist,
00:24:12.13allowing this chromosome to remain
00:24:15.24active in the female cell.
00:24:17.26So, we propose, then, that XX pairing
00:24:21.09is a mechanism of crosstalking
00:24:23.17which allows the two chromosomes
00:24:25.25to adopt mutually exclusive fates,
00:24:28.13of active and inactive X chromosome.
00:24:31.28So, we've also observed that CTCF,
00:24:35.29this very versatile zinc finger transcription factor,
00:24:38.28is essential for X chromosome pairing,
00:24:42.08by serving as an inter-chromosomal glue.
00:24:44.23So, in these complicated experiments,
00:24:47.06what you can see is that
00:24:49.18CTCF binds to Tsix and Xite RNA,
00:24:51.29and CTCF also binds to the DNA
00:24:54.25-- that white line, that 15 kb region I demonstrated before --
00:24:58.15to specific regions of the pairing and choice center.
00:25:04.23So, this binding to the RNA
00:25:07.02is essential for CTCF
00:25:10.04to be recruited as an inter-chromosomal glue.
00:25:12.24So, before concluding this lecture,
00:25:16.12I would like to demonstrate what we think
00:25:20.07is taking place during that process of allelic choice.
00:25:25.20So, prior to the onset of X inactivation,
00:25:28.21this pluripotency factor, OCT4,
00:25:31.24binds to both Tsix and Xite,
00:25:34.13and transactivates the expression of Tsix and Tsix.
00:25:38.18So, I didn't mention that in my lecture,
00:25:40.26but this is the case.
00:25:42.29And then the expression of Tsix and Xite
00:25:45.05recruits CTCF to this pairing region.
00:25:50.15At the onset of X inactivation,
00:25:53.13what we see is that the two chromosomes come together
00:25:56.24and pair exclusively through this 15 kilobase region.
00:26:01.08And we believe that this pairing event
00:26:03.22serves as a platform
00:26:06.19on which the two X chromosomes can communicate with each other,
00:26:09.12and make the determination of
00:26:12.23who will be the active versus the inactive x chromosome.
00:26:15.18Now, exactly what they're saying to each other
00:26:17.27and how this is done
00:26:19.21is something that's under very active investigation.
00:26:22.19We do not presently know.
00:26:24.12However, we suspect that what happens is that
00:26:27.23when the two chromosomes come apart again,
00:26:30.16these transcription factors
00:26:33.08-- like CTCF, and very likely many other factors --
00:26:35.11will repartition onto one X chromosome.
00:26:38.26And so CTCF is serving as a transcriptional repressor.
00:26:42.29Its binding to this chromosome
00:26:46.06will downregulate expression of Tsix as well as Xite.
00:26:51.04And their downregulation is what allows
00:26:53.23Xist to be upregulated from that chromosome.
00:26:57.08And that chromosome becomes the inactive X.
00:27:00.06But on the other hand,
00:27:01.29on the future active chromosome,
00:27:03.15Tsix and Xite persist,
00:27:05.13and their persistence
00:27:08.28prevents Xist from becoming upregulated,
00:27:10.15and that chromosome remains an active chromosome.
00:27:15.28Alright.
00:27:17.20So, before concluding Lecture 2, then,
00:27:20.05I just want to mention one last thing,
00:27:22.20which is that the ends of the sex chromosomes
00:27:27.00-- the telomeres --
00:27:29.11play a very important role during XX pairing.
00:27:33.07Now, XX pairing is not taking place
00:27:35.27in a random place in the nucleus.
00:27:38.00Instead, it's taking place within what we call a tetrad, okay?
00:27:42.16So, what we've now shown is that
00:27:45.04the ends of both sex chromosomes
00:27:46.29-- the X and Y --
00:27:49.26produce a non-coding RNA called PAR-TERRA.
00:27:53.12And PAR-TERRA agglomerates...
00:27:58.03this RNA brings the two telomeres together...
00:28:00.01both X chromosomes,
00:28:01.15and even the X and Y chromosomes.
00:28:02.26It brings the two sex chromosomes together
00:28:05.02at the nuclear envelope.
00:28:06.14And then that RNA serves as a tether,
00:28:08.12and reels in the X inactivation center
00:28:11.28so that pairing takes place within this tetrad
00:28:15.19of two telomeres and two inactivation centers.
00:28:20.13And you can see real-life examples, here,
00:28:22.24by DNA FISH,
00:28:24.20where a pair of the telomeres, shown in red,
00:28:29.07and a pair of inactivation centers, shown in green,
00:28:32.07have agglomerated at the nuclear envelope
00:28:35.03to enable XX pairing.
00:28:37.17So, why would they even bother to do this?
00:28:39.17Well, because the nucleus is a vast space.
00:28:43.04And it would take time for the two inactivation centers...
00:28:45.24a lot of time for the two inactivation centers to come together.
00:28:49.05And so this tethering mechanism
00:28:52.23facilitates this hom*ology searching process
00:28:56.11through this process that we called a constrained diffusion
00:28:58.19in 3-dimensional space.
00:29:01.04So, that then concludes the second lecture.
00:29:05.21And we will talk about
00:29:08.20the initiation and spreading of X inactivation in Lecture 3.

00:00:15.08Hello, everyone.
00:00:16.16Welcome to Lecture 3.
00:00:18.08And again, I'm Jeanie Lee.
00:00:19.25I'm a Professor of Genetics at Harvard Medical School.
00:00:22.07I'm also a faculty member in the Department of Molecular Biology
00:00:25.00at the Massachusetts General Hospital.
00:00:28.17Alright.
00:00:30.05So, we talked about the different steps of X chromosome inactivation,
00:00:34.06and in Lecture 2, we covered, in some detail,
00:00:36.27the molecular mechanisms behind X chromosome
00:00:39.29counting and allelic choice.
00:00:41.25And what I'd like to do in this last lecture
00:00:44.25is to talk about the step of nucleation,
00:00:48.11followed by the spreading of silencing
00:00:51.02across the rest of the X chromosome.
00:00:53.10And recall that the problem is that...
00:00:56.18here we have Xist RNA,
00:00:58.07which is shown in red.
00:01:00.04It coats the inactive X chromosome in cis.
00:01:02.17And it does so without spilling over
00:01:05.01into any of the other X chromosomes...
00:01:07.26into any of the other X chromosomes in the cell,
00:01:10.04or to any autosomes for that matter.
00:01:13.08And so a question is,
00:01:14.29how does this RNA stay put on that chromosome?
00:01:18.06And we know that this property belongs to
00:01:21.12the small piece of DNA called the X inactivation center,
00:01:24.13because when we place the inactivation center
00:01:27.09on an autosome -- so, a non-sex chromosome --
00:01:30.01we see that Xist can also coat that chromosome,
00:01:32.28in cis,
00:01:34.12without spilling over into any of the other chromosomes,
00:01:36.17including the X chromosome
00:01:39.03where it normally would be produced and would be spread.
00:01:42.08Okay.
00:01:43.24So, we're gonna start with this concept of nucleation.
00:01:47.03Now, this is something that we stumbled upon sort of by accident.
00:01:51.16So, recall this experiment,
00:01:53.17which we had done back in the 1990s,
00:01:55.23that sort of led to this dogma that Xist RNA
00:01:59.08is strictly cis-acting.
00:02:01.11So, the idea here was to place an inactivation center,
00:02:04.18or Xist,
00:02:06.13onto an autosome in a male cell,
00:02:08.06and we could observe that Xist
00:02:10.28is produced from this ectopic inactivation center,
00:02:13.29and that that RNA spreads strictly in cis
00:02:17.27across the autosome, without ever diffusing,
00:02:22.19or so we thought, to the natural X chromosome,
00:02:26.27which is located in a different place in the genome,
00:02:29.07or in the nucleus.
00:02:31.01Okay.
00:02:32.19So, that has been recapitulated
00:02:35.11many times by various studies.
00:02:37.27But then, many years later,
00:02:39.20when we repeated the same experiment
00:02:42.00-- except this time we did it in female cells --
00:02:45.00and in the post-inactivation state,
00:02:49.02we got a very surprising finding.
00:02:52.16So, now here are two X chromosomes.
00:02:54.29One is inactive and being coated by Xist RNA.
00:02:58.11We place an inactivation center, or Xist,
00:03:01.20on an autosome.
00:03:03.03And we saw, as we expected,
00:03:05.09that Xist would be upregulated and would coat and silence
00:03:09.10that chromosome in cis.
00:03:12.05However, what we didn't anticipate was that
00:03:15.15within 24 to 48 hours of doing this
00:03:19.19Xist RNA started to fade away from the inactive X.
00:03:24.20Now, upon probing deeper,
00:03:26.26we realized that Xist wasn't actually going away or disappearing.
00:03:30.29In fact, what was happening to it was that it was
00:03:33.27being pulled away onto the autosome,
00:03:37.20which had this multimerized X inactivation center.
00:03:42.22So, that was extremely puzzling.
00:03:45.17And here's the actual experiment,
00:03:47.10where you can see the autosome
00:03:50.06very nicely producing Xist and the X chromosome, down here,
00:03:53.25which is starting to fade away,
00:03:56.04at least with respect to Xist RNA.
00:03:58.13And within 24 hours,
00:04:00.15we see the transgenic Xist spot,
00:04:02.24but the X chromosome spot was nowhere to be found.
00:04:06.26So, that was indeed very surprising.
00:04:09.06And what we learned from these experiments
00:04:11.16is that, in fact, contrary to what we were expecting,
00:04:15.11Xist RNA can diffuse in the nucleus.
00:04:19.02But we normally see it...
00:04:21.02see it localizing in cis,
00:04:23.13because there isn't another site in the nucleus
00:04:27.16that would be receptive to Xist.
00:04:30.11Okay.
00:04:32.01So, Xist can diffuse in trans
00:04:34.01and act on a different chromosome.
00:04:35.20So, this led to a revelation
00:04:39.12that there has to be a sort of a nucleation center
00:04:42.07within the Xist gene itself,
00:04:45.02and that these autosomal transgenes
00:04:47.16must somehow contain such a nucleation site.
00:04:51.24And so we went ahead and dug around
00:04:55.25to try to pinpoint this nucleation site,
00:04:59.03and identified three binding sites
00:05:02.02for a transcription factor called YY1
00:05:05.09that was immensely important to this process of nucleation.
00:05:10.27And so, what we showed is that YY1 protein
00:05:13.20binds to these three sites
00:05:16.10and serves as a tether for Xist RNA
00:05:19.03to anchor to the inactive X chromosome.
00:05:22.03And when we mutated these three YY1 binding sites,
00:05:25.12look what happens to Xist.
00:05:27.21It disperses across almost the entire nucleus.
00:05:32.00So, the RNA detaches from the X chromosome
00:05:34.22and floats away into the nucleus.
00:05:37.22Okay.
00:05:39.10So, we could recapitulate that finding
00:05:42.13by depleting the cells of this critical anchor,
00:05:44.29the YY1 protein.
00:05:46.19You see how recapitulates the loss of this Xist focus
00:05:51.11within these female cells.
00:05:53.23And it wasn't because Xist was no longer being produced.
00:05:56.26Xist was still produced,
00:05:59.11but it could no longer attach to this nucleation center.
00:06:03.25Okay.
00:06:05.12So, we conclude that YY1 is a tether for Xist RNA
00:06:08.16at the nucleation site.
00:06:10.14So, this nucleation allows Xist to attach
00:06:15.01prior to the process of spreading.
00:06:17.05It's an absolutely critical initial step
00:06:20.02to X chromosome spreading.
00:06:23.07So, what we observed is that the nucleation site
00:06:26.00is located within or very close to this repeat sequence
00:06:29.27we called repeat F,
00:06:31.27near the 5' end of Xist;
00:06:34.04that it contains a trio of YY1 binding sites;
00:06:36.25the YY1 protein is very important in tethering Xist RNA
00:06:42.05to this nucleation center.
00:06:44.02And so what we imagine is that Xist RNA
00:06:46.14gets produced from this genetic locus
00:06:48.24and it co-transcriptionally loads onto the nucleation site.
00:06:53.27And what we mean by co-transcriptional is,
00:06:56.04as it's getting synthesized,
00:06:57.28when this piece of the RNA gets exposed,
00:07:00.29it detaches...
00:07:03.25attaches to the inactivation center
00:07:07.02before the transcription is complete.
00:07:10.24Okay.
00:07:12.16And then from the nucleation site,
00:07:14.20Xist spreads in three dimensions
00:07:17.01across the rest of the X chromosome.
00:07:20.19Alright.
00:07:21.26So, now let's talk about this concept of spreading itself.
00:07:25.06How does that actually happen?
00:07:28.01Alright.
00:07:30.00So, what happens from here,
00:07:31.27which is the synthesis of Xist
00:07:34.01and attachment to the nucleation site,
00:07:36.15all the way down to here,
00:07:38.01which is a chromosome condensation
00:07:41.16and a lockdown of gene expression from that chromosome,
00:07:43.08remains largely unknown.
00:07:44.29However, we do know that several things
00:07:47.19have to happen in between.
00:07:49.23One of which is that Xist has to
00:07:52.29push away these factors that normally give rise to gene expression
00:07:56.13-- so, the activating factors --
00:07:58.16and at the same time it has to
00:08:01.16recruit repressive factors to the X chromosome
00:08:05.14to start the process of silencing.
00:08:08.13And then, probably at the same time,
00:08:12.01it is generating a new kind of topology
00:08:14.15on the X chromosome.
00:08:16.09Alright.
00:08:17.28So, how does all of this work.
00:08:19.23Well, we know that Xist is multifunctional.
00:08:22.03And to identify important domains
00:08:24.22that are essential for spreading,
00:08:27.02we have performed a systematic deletional analysis of Xist
00:08:31.10-- so, here again is Xist --
00:08:33.06and what we've done is use the CRISPR/Cas9 gene editing technology
00:08:37.26to remove, sequentially,
00:08:40.121-2 kilobase regions from Xist RNA.
00:08:44.12And we do this at the endogenous Xist locus
00:08:47.25in female cells,
00:08:49.21so that we can observe everything happening in the normal physiological state.
00:08:54.23And what we found was...
00:08:57.14are two repeats that are very important to this process of spreading.
00:09:02.00Okay.
00:09:03.13So, there's repeat B, shown right there,
00:09:05.11and repeat E, which is in the last exon of Xist.
00:09:10.29So normally, as you know by now,
00:09:12.27Xist forms this very tight focus
00:09:15.28over the inactive X chromosome.
00:09:18.06But when we deleted either repeat B or repeat E,
00:09:22.21what you see is a dispersal of the Xist RNA cloud,
00:09:27.21sort of across the nucleus.
00:09:30.13Okay, so I think that's very nicely illustrated here
00:09:32.29and here,
00:09:34.13relative to the wildtype Xist RNA.
00:09:38.07So, to get a better understanding of
00:09:41.18what's happening at the molecular level,
00:09:43.11we performed an epigenetic method called CHART-seq,
00:09:48.03and this allows us to map RNA binding sites on chromatin.
00:09:51.18In this case, we were interested in Xist RNA.
00:09:54.27So, here I'm showing you a time course analysis
00:09:57.19of Xist binding during X inactivation.
00:10:00.16And in this top row, here,
00:10:02.12you can see that Xist is
00:10:05.04initially expressed from the X inactivation center.
00:10:07.06It's nucleating there.
00:10:08.28But then it spreads sort of all at once
00:10:11.08to the rest of the X chromosome.
00:10:14.04In other words, it's not spreading sort of...
00:10:18.04sort of locus by locus,
00:10:20.10in two dimensions down the chromosome.
00:10:21.26But instead, it's spreading
00:10:24.01all at once in three dimensions,
00:10:25.22because you're seeing that the RNA
00:10:28.00is piling up across the X chromosome
00:10:29.24more or less at the same time.
00:10:31.14And the other thing we found out
00:10:33.29from this experiment is that Xist is
00:10:37.14first going to gene-rich regions,
00:10:39.03and in particular it's going to active genes,
00:10:41.19exactly the genes that it ought to be attacking first
00:10:45.26if we're talking about silencing of an entire chromosome.
00:10:49.20And then, eventually,
00:10:52.00Xist spreads over the entire X chromosome,
00:10:54.11until it essentially covers
00:10:57.12the entire 166 megabase chromosome.
00:11:00.28Okay.
00:11:02.10So, we believe that Xist RNA spreads in three dimensions
00:11:05.09across the inactive X chromosome.
00:11:06.27It first nucleates here, at the X inactivation center,
00:11:10.02and then the RNA is transferred through proximity
00:11:13.25to various secondary sites,
00:11:16.24of which there are probably about 100 or so,
00:11:20.24first targeting the gene-rich,
00:11:23.06or the active gene regions,
00:11:25.04before spreading to the rest of the X chromosome.
00:11:28.19Alright.
00:11:29.22So, what happens when we delete this critical repeat B,
00:11:32.16which is important for spreading?
00:11:35.07So, again, here in the top row,
00:11:37.06we can see the wildtype spreading pattern
00:11:39.05that's covering essentially the entire chromosome.
00:11:41.29But when we delete repeat B,
00:11:43.19you see that there is a diminution of binding
00:11:47.07across the entire X.
00:11:49.21And it appears as though
00:11:54.01the ends of the X chromosome are being more affected...
00:11:57.05are more affected than the middle region,
00:12:00.16consistent with this idea of a spreading defect
00:12:03.22in three dimensions.
00:12:06.19Okay.
00:12:08.21And as you would expect of an RNA
00:12:11.07that can no longer spread efficiently on the X chromosome,
00:12:14.14gene silencing is severely compromised.
00:12:17.21So here, by RNA sequencing,
00:12:20.02you can see lots of activity still
00:12:22.29across these two representative genes.
00:12:24.24And the activity is shown here
00:12:26.16as these little red tick marks.
00:12:28.09And there is almost as much activity on the inactive X chromosome
00:12:31.24as on the active X chromosome.
00:12:35.11Okay.
00:12:36.24So now, we've also come to understand,
00:12:39.07in spite of what I just showed you by CHART-sequencing,
00:12:42.26that Xist covers the whole chromosome...
00:12:45.19when we looked at individual cells,
00:12:48.09okay?,
00:12:49.16by super-resolution imaging with a resolution of 20 nanometers,
00:12:53.08we see that Xist isn't actually plastered
00:12:56.24on the entire chromosome.
00:12:58.06It's not really a coat.
00:12:59.26It's actually a cluster of about 100 dots,
00:13:03.23with each dot representing 1-2 Xist transcripts,
00:13:08.22or 1-2 Xist molecules.
00:13:10.28Alright.
00:13:12.16So then, with only 100-200 Xist particles
00:13:15.09on the inactive X chromosome,
00:13:17.09if we were to place the Xist particles end-on-end
00:13:21.03there would only be enough Xist
00:13:23.09to cover about 1% of this 160 megabase chromosome.
00:13:27.20Or put differently,
00:13:30.07there's only one Xist molecule for every 10-20 genes.
00:13:34.08So, that raises this question,
00:13:36.11because Xist is at a stoichiometric disadvantage,
00:13:39.06how does it actually silence an entire chromosome?
00:13:43.01And with respect to that,
00:13:45.14you see that there's sort of a discrepancy
00:13:48.05between what we're seeing at the single cell level
00:13:50.23by super-resolution imaging
00:13:52.20and what we're seeing by CHART-sequencing,
00:13:54.27which really is measuring a population average
00:13:58.10across millions of cells.
00:14:00.14And so, between what we're seeing
00:14:02.11in a single snapshot in time,
00:14:04.11by CHART,
00:14:05.23and what we're seeing at the single cell level,
00:14:08.02we can deduce that Xist is actively moving around,
00:14:11.22very dynamically moving around the X chromosome.
00:14:15.08So, that relates to its question of stoichiometry.
00:14:20.02So, it turns out that Xist
00:14:23.08overcomes this unfavorable stoichiometry
00:14:25.26by recruiting catalytic factors.
00:14:28.21And these are factors that can amplify the work of Xist.
00:14:32.27So, for example,
00:14:35.00here I'm showing four different catalytic factors
00:14:38.06that Xist is directly interacting with,
00:14:41.19including chromosome architectural factors,
00:14:44.02the cohesins,
00:14:45.21the SWI/SNF factors,
00:14:47.24polycomb repressive complexes,
00:14:50.02as well as this non-canonical SMC protein
00:14:53.29called SMCHD1.
00:14:57.04And you'll note from this list of interacting proteins
00:14:59.19that Xist is both coming in contact
00:15:03.07with activators as well as repressors.
00:15:06.17And that is because, in fact,
00:15:08.21during the process of spreading
00:15:10.28it is interacting with both the activating factors
00:15:14.27as well as the repressing factors.
00:15:18.11So, we now turn our attention
00:15:20.22to the first function of Xist,
00:15:22.10which is the recruitment of repressive factors.
00:15:25.12And one of the factors that it recruits
00:15:28.07is this PRC2, or polycomb repressive complex 2.
00:15:32.05This is an epigenetic complex
00:15:34.28that trimethylates histone H3 at lysine 27.
00:15:39.23And it's a very important enzyme that represses gene expression.
00:15:44.03And it's important throughout development,
00:15:46.11as well as during the etiology of disease.
00:15:50.10However, there has been a long-standing question in the field
00:15:54.03-- not just about polycomb complexes
00:15:56.29but about many, many other epigenetic complexes --
00:16:00.08about how they can be targeted to specific locations in our genome
00:16:07.06when these complexes are largely devoid
00:16:10.24of a sequence-specific DNA binding subunit.
00:16:13.08Okay.
00:16:14.20So, how do they know where to go?
00:16:16.17Now, the answer to this question is going to be multifaceted, of course.
00:16:19.15There are gonna be many different recruiting mechanisms,
00:16:21.17including transcription factors,
00:16:23.06including specific motifs in DNA and whatnot.
00:16:28.05But we believe that a major piece of the puzzle
00:16:31.27lies in the non-coding RNA,
00:16:34.19and sometimes even coding RNA,
00:16:36.25dating back to an experiment that we did 10 years ago,
00:16:40.22in which we demonstrated that
00:16:43.12PRC2 can directly interact with RNA,
00:16:47.13in this case Xist,
00:16:49.13and recruit PRC2 to the X chromosome.
00:16:53.29And it's doing so through a motif
00:16:56.28at the very 5' prime end of Xist called repeat A.
00:17:01.14So, this is a biochemical analysis
00:17:03.23that shows that PRC2 interacts with Xist RNA
00:17:07.09with high affinity,
00:17:09.02with a dissociation constant of 20-80 nanomolar,
00:17:12.11which is a very good dissociation constant
00:17:14.16for an RNA binding protein.
00:17:16.11And it contrasts with the affinities...
00:17:19.00some very low affinities for these nonspecific RNAs
00:17:22.06from various other species, like Tetrahymena and bacteria.
00:17:26.06Okay.
00:17:28.01So, now in this slide
00:17:30.07I will attempt to convey the complex dynamics
00:17:33.16that occur between the RNA and PRC2.
00:17:37.24So, as we envision it...
00:17:40.10now, initially, the RNA that contains this repeat A motif
00:17:44.13will attract polycomb repressive complexes
00:17:47.24to the X inactivation center.
00:17:49.21Okay.
00:17:51.07Now, very importantly,
00:17:53.03our genetic and biochemical experiments
00:17:55.06show that even though the long non-coding RNA
00:17:58.15is recruiting PRC2 in a site-specific manner,
00:18:03.10that does not mean that it is
00:18:07.03automatically going to load that complex onto chromatin,
00:18:10.23or that it will induce the catalysis on H3K27.
00:18:16.01Okay.
00:18:17.02So, in fact, these steps are biochemically
00:18:19.14and genetically separable.
00:18:20.27So, for example,
00:18:22.26as long as the antisense repressor, Tsix,
00:18:25.14is expressed from the inactivation center,
00:18:28.06we see that this complex
00:18:30.19does not load onto chromatin.
00:18:32.14So, at this time we can see a RIP,
00:18:35.10RNA immunoprecipitation,
00:18:37.13between PRC2 and repeat A RNA.
00:18:40.02But we do not see them ChIPing
00:18:42.20onto the 5' end of Xist.
00:18:47.00And it is only when the antisense RNA disappears
00:18:50.15that we see that complex load onto the X chromosome.
00:18:54.09But even so, that does not by itself
00:18:57.20unleash the methyltransferase activity
00:19:00.08of the catalytic subunit, EZH2.
00:19:02.27However, when this complex comes into contact
00:19:06.02with this accessory subunit,
00:19:07.26called JARID2,
00:19:09.19we see that the affinity of the RNA
00:19:13.22for the catalytic subunit, EZH2,
00:19:16.08decreases.
00:19:18.06So, the affinity goes down.
00:19:20.07And that loss of binding of the RNA to EZH2
00:19:24.23is associated with unleashing of
00:19:28.23the histone methyltransferase activity.
00:19:30.26So, again, this principle illustrates
00:19:33.21how we can separate polycomb recruitment
00:19:37.14from its loading onto chromatin
00:19:39.19to its catalysis on H3K27.
00:19:43.13Okay.
00:19:45.08So, the take-home message...
00:19:47.13long non-coding RNAs can recruit PRC2,
00:19:49.29but hold their activity...
00:19:52.15or, hold the activity of PRC2 in check
00:19:54.28until a developmental signal is received,
00:19:58.01at which time PRC2 and JARID2
00:20:02.12interact to trimethylate histone H3 at lysine 27.
00:20:08.02Xist overcomes its unfavorable stoichiometry
00:20:10.29by recruiting these catalytic factors.
00:20:13.20And we envision that these factors
00:20:16.13use a hit and run mechanism
00:20:18.24to silence the entire chromosome in an efficient manner.
00:20:22.10So, again, Xist first binds to the inactivation center,
00:20:26.19nucleates at that site,
00:20:28.16recruits all these factors,
00:20:30.08and then spreads in three dimensions
00:20:32.27to about 100 sites located across the chromosome.
00:20:35.28And then, at these secondary sites,
00:20:38.08and we're taking as an example, here, PRC2,
00:20:40.25but there are many other catalytic factors...
00:20:42.19so, PRC2 lands,
00:20:45.18and it processively methylates successive nucleosomes
00:20:50.11until it covers about 1 or 2 megabases of chromatin.
00:20:53.28And imagine that this happens
00:20:56.25100 times over across the X chromosome,
00:20:59.07more or less at the same time,
00:21:00.29and then you can see how this process of silencing
00:21:03.24can be amplified
00:21:06.05and can take place in a very efficient manner.
00:21:09.17So, now I'd like to turn your attention
00:21:11.28to the second aspect of Xist function,
00:21:14.10and that is its antagonization or its repulsion
00:21:17.27of the activating factors.
00:21:19.29And we're going to use as an example
00:21:22.13this epigenetic factor called SWI/SNF.
00:21:26.08Okay.
00:21:27.28So, SWI/SNF is an ATP-dependent
00:21:31.16chromatin remodeling enzyme.
00:21:33.09And central to its activity
00:21:35.23is the catalytic subunit, BRG1.
00:21:38.19So, SWI/SNF is normally associated
00:21:41.03with open chromatin.
00:21:43.07Indeed, it makes chromatin accessible,
00:21:46.13poising it for gene activation.
00:21:49.12And so, normally you would find SWI/SNF
00:21:51.29on the active X chromosome
00:21:53.16but not on the inactive X chromosome.
00:21:56.11As Xist RNA spreads over the inactive X chromosome,
00:21:59.14the RNA comes into contact with BRG1
00:22:02.13and inhibits the ATP...
00:22:05.12ATPase activity of BRG1,
00:22:08.09just as it inhibits the methyltransferase activity
00:22:11.19of PRC2.
00:22:13.28So, this is an immunofluorescence experiment,
00:22:16.03and you can see that BRG1
00:22:18.18is normally present throughout the nucleus.
00:22:21.03But where there is Xist RNA,
00:22:23.02shown here in red,
00:22:24.21you can see that there is a depletion of BRG1
00:22:28.14over that same chromosome territory,
00:22:32.02suggesting that Xist RNA evicts BRG...
00:22:36.10BRG1 from the inactive X chromosome.
00:22:39.11Alright.
00:22:40.20So then, we'll now turn to
00:22:43.05the third and final function of Xist,
00:22:45.21and that is its role in directing changes
00:22:49.09in three-dimensional chromosome architecture.
00:22:52.07So, our RNA proteomic analysis
00:22:55.19also showed that Xist is interacting with
00:22:59.05a number of chromosome architectural factors.
00:23:01.25So, you can see here various cohesions,
00:23:04.07as well as CTCF.
00:23:06.24These are two architectural factors
00:23:09.00that go to construct 3D chromatin.
00:23:11.04So, we've already talked extensively about CTCF,
00:23:13.13a zinc finger protein that brings together distant genetic elements
00:23:17.04that form these chromatin loops.
00:23:19.01And then we have cohesins,
00:23:20.20which are this multi-subunit complex
00:23:22.26that forms a ring around the base of the loops
00:23:25.12to lock in that architectural structure.
00:23:29.09So, it's... it is known that
00:23:32.21mammalian chromosomes are organized into two distinct entities,
00:23:36.27one called the topologically associating domain,
00:23:39.08or the TAD,
00:23:41.01and the other the compartment.
00:23:43.11So, TADs are these large loops of chromatin
00:23:46.07of around 1-2 megabases in size,
00:23:49.01within which genes can be,
00:23:51.09though they don't have to be,
00:23:53.07coordinately regulated.
00:23:54.25And then the active TADs,
00:23:56.23or those active loops,
00:23:58.29coalesce and form a separate compartment
00:24:02.11called the A compartment,
00:24:04.16whereas the inactive or the less active genes
00:24:08.28form another type of compartment
00:24:11.13that's called a B compartment.
00:24:13.11So, as you might imagine,
00:24:15.12the inactive X chromosome is organized completely differently
00:24:19.18from all other chromosomes.
00:24:21.01So now, here's an inactive X chromosome.
00:24:23.06The entire chromosome is shown across the top.
00:24:25.22And what I've done here is
00:24:27.22magnified the two ends of the chromosome.
00:24:30.04And you can see from this contact heat map
00:24:32.25that these TADs,
00:24:34.17which are these triangular structures,
00:24:36.20can be seen essentially all across the chromosome.
00:24:39.27So, the active X chromosome
00:24:42.16looks a lot like any other chromosome,
00:24:44.15like all autosomes.
00:24:46.15On the X chromosome, there are about 110
00:24:48.22of these topologically associating domains.
00:24:52.16Now, contrast that with the inactive X chromosome,
00:24:55.21where, yes, you might still be able to
00:24:59.17envision a formation of these topologically associating domains,
00:25:04.05but they are much, much weakened.
00:25:06.15And so, one of the things that Xist has to do
00:25:10.01as it's spreading across the chromosome
00:25:11.22is to attenuate the formation...
00:25:14.03although not abolish...
00:25:15.28attenuate the formation of these so-called TADs.
00:25:19.24When we mutate repeat B
00:25:23.00-- so, that's the critical domain for X...
00:25:26.08for Xist spreading --
00:25:27.24we see that nothing happens to the active X chromosome.
00:25:31.04The active X chromosome still has
00:25:33.07110 or so TADs.
00:25:35.21But on the other hand,
00:25:37.24when we do the same thing to the inactive X chromosome,
00:25:40.17you start to see that these TADs
00:25:43.07persist on that chromosome,
00:25:45.01or that maybe they're even coming back.
00:25:47.06They either persist or come back,
00:25:49.03suggesting that Xist is very important
00:25:51.25to the attenuation of these topologically associating domains.
00:25:56.07So, how does Xist do this?
00:25:58.02Well, it's probably doing a number of different things,
00:26:01.07one of which is that it is also evicting cohesins
00:26:04.23from that chromosome.
00:26:07.00So, this is one of the subunits of cohesin.
00:26:09.11It's an immunofluorescence
00:26:11.17that shows that cohesins are normally widely distributed
00:26:13.29throughout the nucleus.
00:26:15.13But again, where there is Xist,
00:26:18.10shown here in red,
00:26:20.13you see a depletion of cohesins next to this arrow.
00:26:25.27And the same is true of CTCF.
00:26:29.02Okay.
00:26:30.17Now, when we delete Xist,
00:26:31.29the cohesins come back,
00:26:33.19as shown here by this red peak of cohesins here and here.
00:26:38.17And you can see that in the wildtype chromosome,
00:26:40.29those two red peaks are not present.
00:26:43.23Xist this not only attenuating TADs,
00:26:46.19but it's also directing the formation of these inactive X-specific compartments.
00:26:51.14So, SMD... SMCHD1 is very important
00:26:54.24in the formation of these XI-specific compartments.
00:26:59.01And it was Emma Whitelaw who observed, many years ago,
00:27:02.23that embryos lacking SMCHD1
00:27:05.19would die in mid-gestation
00:27:08.08due to dysfunctional X inactivation.
00:27:10.17So, SMCHD1 is a non-canonical SMC protein
00:27:14.03that's like the cohesins and condensins,
00:27:16.10except that it has a very different function.
00:27:19.07So, here you can see that
00:27:21.24Xist plays a very active role in the recruitment
00:27:24.21and the enrichment of SMCHD1
00:27:26.13along the inactive X chromosome.
00:27:30.00And it's in fact one of the proteins
00:27:32.23that we identified when we performed the Xist proteomic analysis,
00:27:36.09as a factor that directly interacts with Xist.
00:27:39.19So, it turns out that SMCHD1
00:27:42.21plays at least two important roles
00:27:45.14during X inactivation.
00:27:47.00So, in the first, it's aiding the local spreading
00:27:49.25of Xist-PRC2 complexes.
00:27:51.19And in the second,
00:27:53.18it's merging these inactive X-specific compartments.
00:27:57.25Okay.
00:27:59.22So, here's the first role.
00:28:01.15SMCHD1 is important for the regional spreading of Xist.
00:28:05.04And you can see in this epigenomic analysis,
00:28:08.09in the first track,
00:28:10.00that Xist spreads along the...
00:28:12.16this region of the X chromosome
00:28:14.23-- it's about 1 megabase --
00:28:17.04more or less evenly.
00:28:19.09But in the SMCHD1 knockout,
00:28:21.10which is shown in the second track,
00:28:23.04you can see that there is a depletion of Xist
00:28:25.19across this 1 megabase domain,
00:28:28.20which is also green shaded.
00:28:31.02And associated with that
00:28:34.16is a defect in polycomb spreading and H3K27 methylation
00:28:38.07across that same domain.
00:28:41.13So, SMCHD1 is very important
00:28:44.19for regional spreading of Xist.
00:28:46.03And the loss of SM...
00:28:47.25of SMCHD1 results in a defect of spreading,
00:28:51.28as well as a focal loss of H3K27 methylation.
00:28:57.21Now, the other thing that SMCHD1 does
00:29:00.14is that it merges these inactive X-specific compartments.
00:29:04.24So, shown here are Hi-C experiments,
00:29:08.23and what I'm showing is a contact heat map
00:29:12.01for the active X chromosome
00:29:14.11in the wildtype state,
00:29:16.15with the active X chromosome shown here
00:29:18.28along the diagonal.
00:29:20.23And what you probably cannot see at this magnification
00:29:24.28is that there are about 110 topologically associated domains
00:29:28.24on that active chromosome.
00:29:30.21Whereas on the inactive chromosome,
00:29:32.24those TADs are significantly weakened.
00:29:36.15And instead of TADs,
00:29:38.23this X chromosome shows
00:29:41.07two large so-called mega domains.
00:29:43.26So, here's one mega domain,
00:29:46.28and here's another mega domain.
00:29:48.28Now, contrast that with what happens
00:29:50.29when we remove SMCHD1.
00:29:53.10And on the active X chromosome,
00:29:54.29nothing happens.
00:29:56.21But on the inactive X chromosome,
00:29:59.02you see that these two mega domains
00:30:01.17start to break up.
00:30:03.15And that can be much better visualized
00:30:05.18by going to the bottom set of panels.
00:30:08.11These are heat maps of
00:30:10.20Pearson correlation coefficients.
00:30:13.12And what you can see is that in the wildtype inactive X chromosome,
00:30:19.13these two mega domains really pop out, as shown in red.
00:30:25.15But when SMCHD1 is removed,
00:30:29.15those two mega domains adopt a checkerboard pattern.
00:30:33.14Okay.
00:30:34.28And the checkerboard pattern that you see on the mutant inactive X
00:30:37.29is distinctly different from what you see
00:30:41.04on the active X chromosome,
00:30:43.03which shows a much finer checkerboarding pattern.
00:30:46.08That's consistent with the A/B compartments
00:30:49.06that we talked about a few slides ago.
00:30:52.24So, we can better visualize this
00:30:55.03by going to a principal component analysis.
00:30:57.21And in the first principal component,
00:30:59.19you can see these red/blue structures
00:31:02.21on the active X chromosome,
00:31:04.25which are the A/B compartments.
00:31:08.19And when we remove SMCHD1,
00:31:11.10nothing happens to the active X chromosome.
00:31:14.20It's impervious to there being a loss of SMCHD1.
00:31:19.08But then here's the act...
00:31:21.04the inactive X chromosome.
00:31:23.09And you can see that in the wildtype state
00:31:25.29there are these two megadomains,
00:31:27.20one in blue and one in red.
00:31:29.23But when we remove SMCHD1,
00:31:32.02those two megadomains break up
00:31:35.07into these finer structures that we call
00:31:38.01S1 and S2 compartments,
00:31:40.22referring to the fact that they appear only when we remove SMCHD1.
00:31:46.00But also appreciate that these finer structure...
00:31:49.02these finer structures are not the A/B compartments
00:31:53.22that you see on the active X chromosome.
00:31:57.08So, it turns out that these S1/S2 compartments
00:32:00.18are intermediate structures
00:32:03.01during the formation of the inactive X.
00:32:06.04So, prior to X inactivation,
00:32:08.12we see these A/B compartments,
00:32:10.09like we see on all chromosomes.
00:32:12.28But then at the onset of X inactivation,
00:32:15.25as Xist spreads over the X chromosome,
00:32:18.12it merges...
00:32:22.09it being Xist...
00:32:24.03merges the red and blue compartments
00:32:27.11to form these S1/S2 structures,
00:32:30.18these larger S1/S2 structures.
00:32:34.10And that's... as X inactivation proceeds,
00:32:36.27Xist recruits SMCHD1,
00:32:40.08and SMCHD1 in turn
00:32:43.20merges the S1/S2 compartments
00:32:47.09into these two megadomains
00:32:50.05to form a compartment-less chromosome.
00:32:52.26Alright.
00:32:54.12So, in these three lectures
00:32:56.05I've thrown a lot of information at you.
00:32:57.24And I what I'm going to attempt to do in this final slide
00:33:00.08is to integrate some of that information.
00:33:03.27So, we envision that at the onset of X inactivation
00:33:09.22this RNA with the repeat A motif
00:33:12.02recruits PRC2 to the X inactivation center.
00:33:16.29But as I mentioned, the recruitment process...
00:33:21.00the RNA isn't very important for the recruitment process,
00:33:23.15but it also holds the activity of PRC2 in check.
00:33:28.22So, recruitment is not the same thing as loading,
00:33:31.10which is not the same thing as catalysis.
00:33:33.24Because as long as the antisense RNA is expressed,
00:33:36.23PRC2 is prevented from loading onto chromatin.
00:33:42.01And I also mentioned, in Lecture 2,
00:33:44.27that at this time one of the very first things
00:33:48.12that we see during cell differentiation
00:33:50.27is a pairing of the two X chromosomes.
00:33:53.01And as a result of this pairing process,
00:33:55.13there's a mutually exclusive determination
00:33:58.02of the active and the inactive X chromosome,
00:34:01.05presumably through an asymmetric expression pattern
00:34:06.25of the antisense RNA, Tsix.
00:34:10.05So, from the future inactive X chromosome,
00:34:13.27the antisense RNA disappears,
00:34:16.06while the antisense RNA
00:34:19.06persists on the future active X chromosome.
00:34:22.17Alright.
00:34:24.14So, then on the future inactive X chromosome,
00:34:26.07the disappearance of the antisense RNA
00:34:29.01allows PRC2 to load onto chromatin.
00:34:31.29But again, that is the not...
00:34:34.12not enough to unleash the methyltransferase activity of EZH2.
00:34:40.10When the RNA comes into contact with the accessory subunit JARID2,
00:34:44.15the affinity of the RNA for PRC2 decreases,
00:34:49.20the RNA is at least partially dislodged,
00:34:53.17and that unleashes the methyltransferase activity of EZH2.
00:34:58.19Okay.
00:35:00.17So, while all of this is happening,
00:35:03.25at the same time we see that the Jpx RNA
00:35:07.09-- this RNA which is just upstream of Xist --
00:35:09.16is transcriptionally upregulated tenfold.
00:35:13.23And when it crosses a certain threshold,
00:35:16.11as it will do only in the female cell,
00:35:19.05because the female has two copies of Jpx,
00:35:21.16Jpx evicts CTCF from the 5' end of Xist.
00:35:30.11And at probably around the same time,
00:35:33.24the Gribnau lab has shown that this transcriptional repressor, REXI,
00:35:38.06is degraded by an E3-ubiquitin ligase called Rnf12.
00:35:43.26And it's really the combination of all of these events
00:35:49.10-- in particular the downregulation of the antisense RNA,
00:35:54.04the upregulation of Jpx RNA,
00:35:56.02and the eviction of CTC...
00:35:57.19CTCF and REXI --
00:35:59.27that allows full-length Xist to be expressed
00:36:04.20for the first time.
00:36:06.11So, Xist of course
00:36:08.29also has a binding site for PRC2.
00:36:11.03And we now know from RNA proteomic analysis
00:36:14.01that Xist probably binds to about
00:36:16.11100 other proteins.
00:36:18.11Now, this RNA protein complex
00:36:20.10has to first attach, or load,
00:36:24.09onto a single nucleation site
00:36:28.21through YY1, the transcription factor YY1.
00:36:31.22Now, without attaching to YY1
00:36:33.25this RNA-protein complex will diffuse
00:36:36.29through the rest of the nucleus.
00:36:39.13So, from this single nucleation center,
00:36:42.01the RNA-protein complex then spreads in three dimensions
00:36:46.03across the rest of the X chromosome.
00:36:48.16And again, it does three things.
00:36:51.02It's not only recruiting silencing factors
00:36:53.13to the rest of the X chromosome,
00:36:55.07but it's also evicting activating factors
00:36:58.06like cohesins and BRG1,
00:37:00.24or the SWI/SNF factors,
00:37:02.25from that X chromosome.
00:37:05.04So, that, in a nutshell,
00:37:07.10is how we're viewing the initiation and spreading
00:37:10.08of X inactivation.
00:37:11.22And I should say, by way of conclusion,
00:37:14.02that this is a model.
00:37:16.07So, it is a facsimile of what we can't actually directly visualize
00:37:20.28in the natural world.
00:37:23.04And scientists use these models as basic frameworks,
00:37:27.01within which we can design additional experiments
00:37:29.22to probe, to refute,
00:37:32.21or to accept a hypothesis.
00:37:34.03And of course, scientists disagree all the time with each other
00:37:37.07about exactly how things are working in nature.
00:37:39.25And so, we hope that additional data
00:37:42.07that we will generate in the coming years
00:37:44.17will allow us to continually refine this model.
00:37:48.08And I hope that I'll be able to share
00:37:50.24some of those new ideas with you
00:37:53.01in the coming years.