Xist is often described as ‘painting’ the inactive X and it’s a very good description. Let’s revert yet again to our analogy of the DNA code as a script. This time we’ll imagine that the script is written on a wall, maybe it’s an inspiring poem or speech in a classroom. At the end of the summer term the school building closes down and is sold for conversion to apartments. The decorators arrive and paint over the script. Now there’s nothing to tell the new residents to ‘play up and play the game’, or exactly how they should ‘meet with Triumph and Disaster’. But the instructions are actually still there, they’re just hidden from view.
When Xist binds over the X chromosome that produced it, it induces a kind of creeping epigenetic paralysis. It covers more and more genes, switching them off. It first seems to do this by acting as a barrier between the genes and the enzymes that normally copy them into mRNA. But as the X inactivation gets better established, it changes the epigenetic modifications on the chromosome. The histone modifications that normally turn genes on are removed. They are replaced by repressive histone modifications that turn genes off.
Some of the normal histones are removed altogether. Histone H2A is replaced by a related but subtly different molecule called macroH2A, strongly associated with gene repression. The promoters of genes undergo DNA methylation, an even more stringent way of turning the genes off. All these changes lead to binding of more and more repressor molecules, coating the DNA on the inactive X and making it less and less accessible to the enzymes that transcribe genes. Eventually, the DNA on the X chromosome gets incredibly tightly wound up, like a giant wet towel being turned at each end, and the whole chromosome moves to the edge of the nucleus. By this stage most of the X chromosome is completely inactive, except for the Xist gene, which is a little pool of activity in the midst of a transcriptional desert[108].
Whenever a cell divides, the modifications to the inactive X are copied over from mother cell to daughter cell, and so the same X remains inactivated in all subsequent generations of that starter cell.
While the effects of Xist are amazing, the description above still leaves a lot of questions unanswered. How is Xist expression controlled? Why does it switch on when ES cells start to differentiate? Is Xist only functional when it’s in female cells, or could it act in males cells too?
The last question was first addressed in the lab of Rudi Jaenisch, whom we met in the context of iPS cells and Shinya Yamanaka’s work in Chapter 2. In 1996, Professor Jaenisch and his colleagues created mice carrying a genetically engineered version of the X Inactivation Centre (an X Inactivation Centre transgene). This was 450kb in size, and included the Xist gene plus other sequences on either side. They inserted this into an autosome (non-sex chromosome), created male mice carrying this transgene, and studied ES cells from these mice. The male mice only contained one normal X chromosome, because they have the XY karyotype. However, they had two X Inactivation Centres. One was on the normal X chromosome, and one was on the transgene on the autosome. When the researchers differentiated the ES cells from these mice, they found that Xist could be expressed from either of the X Inactivation Centres. When Xist was expressed, it inactivated the chromosome from which it was expressed, even if this was the autosome carrying the transgene[109].
These experiments showed that even cells that are normally male (XY) can count their X chromosomes. Actually, to be more specific, it showed they could count their X Inactivation Centres. The data also demonstrated that the critical features for counting, choosing and initiation were all present in the 450kb of the X Inactivation Centre around the Xist gene.
We know a bit more now about the mechanism of chromosome counting. Cells don’t normally count their autosomes. Both copies of chromosome 1, for example, operate independently. But we know that the two copies of the X chromosome in a female ES cell somehow communicate with each other. When X inactivation is getting going, the two X chromosomes in a cell do something very weird.
They kiss.
That’s a very anthropomorphic way of describing the event, but it’s a pretty good description. The ‘kiss’ only lasts a couple of hours or so, and it’s startling to think this sets a pattern that can persist in cells for the next hundred years, if a woman lives that long. This chromosomal smooch was first shown in 1996 by Jeannie Lee, who started out as a post-doctoral researcher in Rudi Jaenisch’s lab, but who is now a professor in her own right at Harvard Medical School, where she was one of the youngest professors ever appointed. She showed that essentially the two copies of the X find each other and make physical contact. This physical contact is only over a really small fraction of the whole chromosome, but it’s essential for triggering inactivation[110]. If it doesn’t happen, then the X chromosome assumes it is all alone in the cell, Xist never gets switched on, and there is no X inactivation. This is a key stage in chromosome counting.
It was Jeannie Lee’s lab that also identified one of the critical genes that controls Xist expression[111]. DNA is double-stranded, with the bases in the middle holding the strands together. Although we often envisage it as looking like a railway track, it might be better to think of it as two cable cars, running in opposite directions. If we use this metaphor, then the X Inactivation Centre looks a bit like Figure 9.4.
Figure 9.4 The two strands of DNA at a specific location on the X chromosome can each be copied to create mRNA molecules. The two backbones are copied in opposite directions to each other, allowing the same region of the X chromosome to produce Xist RNA or Tsix RNA.
There is another non-coding RNA, about 40kb in length, in the same stretch of DNA as Xist. It overlaps with Xist but is on the opposite strand of the DNA molecule. It is transcribed into RNA in the opposite direction to Xist and is referred to as an antisense transcript. Its name is Tsix. The eagle-eyed reader will notice that Tsix is Xist backwards, which has an unexpectedly elegant logic to it.
This overlap in location between Tsix and Xist is really significant in terms of how they interact, but it makes it exceedingly tricky to perform conclusive experiments. That’s because it’s very difficult to mutate one of the genes without mutating its partner on the opposite strand, a sort of collateral damage. Despite this, considerable strides have been made in understanding how Tsix influences Xist.
If an X chromosome expresses Tsix, this prevents Xist expression from the same chromosome. Oddly enough, it may be the simple action of transcribing Tsix that prevents the Xist expression, rather than the Tsix ncRNA itself. This is analogous to a mortice lock. If I lock a mortice from the inside of my house and leave the key in the lock, my partner can’t unlock the door from the outside of the house. I don’t need to keep locking the door, just having the key in there is enough to stop the action of someone on the other side. So, when Tsix is switched on, Xist is switched off and the X chromosome is active.
108
For a good review, see Lee, J. T. (2010)