X-chromosomes like to accessorize. Some of them, specifically one in each female cell, are not content with the standard chromatin suit—how boring! Instead, they dress up in an RNA cloak of their own design. This is not just a fashion statement. The production of this RNA, which is noncoding, is essential to the process of X-chromosome inactivation and thus to the survival of the developing female embryo. How does a large noncoding RNA transcript bring about the essential epigenetic changes associated with X-chromosome inactivation? HMS researcher Jeannie Lee is one step closer to answering that question.
Lee, a Howard Hughes investigator and HMS professor of genetics in the Department of Molecular Biology at Massachusetts General Hospital, studies large noncoding RNAs (ncRNAs), a heterogeneous and growing family of RNA molecules distinct from other noncoding RNAs like siRNAs, miRNAs and piRNAs, which are much smaller. Recently overshadowed by their tiny relatives, large ncRNAs are proving to have important functions in gene regulation and chromatin remodeling. “They are going to be just as important as small RNAs in regulating gene expression,” said Lee.
In support of this prediction, Lee and her colleagues have recently discovered a large ncRNA, they call RepA, which is required for X-chromosome inactivation. In addition to upregulating another essential large ncRNA, known as Xist, RepA recruits a complex of polycomb proteins, known to epigenetically silence genes by methylating histones. These results, which suggest that large ncRNAs play a direct role in targeting genes for silencing, are published in the Oct. 31 issue of Science.
Prior to implantation, each cell in a female embryo must inactivate one X-chromosome to equalize gene dosage with XY males. In extra-embryonic tissue, it is always the paternally inherited X-chromosome that is silenced, but for somatic cells, chromosome selection is random and consequential—the inactivated chromosome will be silenced for a lifetime.
Before one X-chromosome in a female cell is epigenetically silenced, certain biochemical alterations begin to distinguish one X-chromosome from the other. Early in the inactivation process, the future inactive X-chromosome begins to produce transcripts of the 17kb ncRNA Xist. Exuberantly transcribed from a central point on the chromosome, Xist transcripts soon spread to the chromosome’s distal tips. Under the microscope, Xist appears as an amorphous cloud literally cloaking the entire chromosome.
In this study, Lee and colleagues describe the mechanism by which this cloak of Xist RNA initiates the silencing of an X-chromosome. “It was known that Xist coats the chromosome and that polycomb group proteins somehow find their way there, but no one knew the relationship between these events,” said first author Jing (Crystal) Zhao, a postdoctoral fellow in Lee’s lab.
This gap in knowledge was not for lack of effort. According to Lee, over the years many research groups, including her own, had attempted to identify proteins that directly interact with Xist, reasoning that Xist might recruit silencing factors. However, the problem seemed intractable until Zhao tried an RNA immunoprecipitation technique and pulled down Xist RNA with a complex of polycomb proteins, called PRC2. Not only were these the first Xist-interacting proteins to be identified, the interaction provided strong evidence that the RNA itself, not the DNA, recruits essential silencing factors to the X-chromosome.
The PRC2 complex, which was first identified in Drosophila, silences genes by stimulating chromatin remodeling, thereby preventing subsequent transcriptional activation, and in the case of the X-chromosome, sealing the fate of the inactive X. The PRC2 complex, which is perhaps best known for silencing developmental control genes such as the HOX genes, consists of multiple, partially interchangeable subunits. One subunit, called Ezh2, catalyzes the trimethylation of histone H3 at lysine 27, leaving behind this repressive epigenetic mark. Lee and Zhao found that Ezh2 is also the RNA-binding subunit, recognizing Xist RNA and pulling the rest of the PRC2 complex to the site of silencing.
The identification of a direct interaction between PRC2 and Xist RNA was exciting, but it presented Lee and her colleagues with a puzzle. They had observed that the trimethylation of histone H3 at lysine 27 occurs before the dramatic upregulation of Xist on the future inactive X-chromosome, implying that the methylating polycomb complex was there first. A clue to this sequence of events came through their observation that the 5-prime end of the Xist transcript accumulates more PRC2 and at an earlier time than the 3-prime end. Ultimately, they determined that a motif at the 5-prime end of Xist, called Repeat A, is the original epicenter of PRC2 binding.
Although Repeat A had been previously shown to be required for X-inactivation, little was known about this motif. Using genetic techniques and reporter assays, the researchers made the surprising discovery that Repeat A is actually an independent transcriptional unit that produces a smaller internal RNA transcript, the one they call RepA. It is produced prior to Xist upregulation and is believed to be required for the subsequent upregulation of the larger ncRNA. What’s more, when the researchers transferred the RepA motif to an autosome in mouse embryonic stem (ES) cells, RepA transcripts recruited PRC2 to the autosome, indicating that RepA alone was sufficient to attract the silencing complex.
But RepA is not always successful at this job. Lee and her colleagues found that male ES cells also express RepA, yet their single X-chromosome is not silenced. If RepA is sufficient to attract PRC2 to the chromosome, what protects the male X-chromosome from being methylated and silenced by PRC2? It turns out that another large ncRNA is involved.
The Lee lab had previously identified a large ncRNA they called Tsix (pronounced “sigh-X”). As its name suggests, Tsix is the antisense version of Xist, and it was found to repress Xist expression. Like RepA and Xist, Tsix can bind the PRC2 complex. This is significant because although Tsix is originally expressed on all X-chromosomes, its downregulation on the future inactive X-chromosome releases Xist from repression.
In the model proposed by Lee, Zhao, and their colleagues, Tsix and RepA compete for PRC2 binding on those chromosomes that will remain active, namely the male X-chromosome and one female X-chromosome. In the presence of both large ncRNAs, Tsix essentially titrates PRC2 away from RepA, preempting the histone methylation that they believe is required for Xist upregulation and subsequent X-chromosome inactivation. In contrast, when Tsix is downregulated on the future inactive X-chromosome early in embryogenesis, RepA attracts the entire pool of PRC2 and thus facilitates chromosomal silencing.
Lee believes that other RNA molecules might also attract polycomb proteins and, perhaps, other chromatin remodeling complexes as well. “I would guess there are other noncoding RNAs out there doing a similar job,” she said, “and we are working to find out what they are.”
For Students: Contact Jeannie Lee at lee@molbio.mgh.harvard.edu for more information on this and other lab projects.
Conflict Disclosure: The authors declare no conflicts of interest
Funding Sources: The Howard Hughes Medical Institute, the National Institutes of Health
