These data enabled researchers to narrow down the region on the X chromosome that was key for inactivation. Rationally enough, they called this region the X inactivation centre. In 1991, a group reported that this region contained a gene that they called Xist.[14] Only the Xist gene on the inactive chromosome expressed Xist RNA.{119},{120} This made perfect sense, because X inactivation is an asymmetric process. In a pair of equivalent X chromosomes, one is inactivated and one is not. So it seemed consistent that this process would be driven by a scenario where one chromosome expresses a gene and the other doesn’t.
It was obvious that the next question would be to ask how Xist works and the first thing that researchers did was to try to predict the sequence of the protein that it produced. This is usually relatively straightforward. Once they had found the sequence of the Xist RNA molecule, all that the scientists had to do was run this through a simple computer program that would predict the encoded amino acid sequence. Xist RNA is very long, about 17,000 bases. Each amino acid is encoded by a block of three bases, so a 17,000-base RNA could theoretically code for a protein of over 5,700 amino acids. But when the Xist RNA sequence was examined, the longest run of amino acids was just under 300. This was despite the fact that the Xist RNA was spliced, in the way we first saw in Chapter 2, so had lost all the intervening junk sequences.
The ‘problem’ was that the Xist RNA was liberally scattered with sequences that don’t code for amino acids, but which act as stop signals when protein chains are being built up. We can envisage this as being a little like trying to build a tall tower out of LEGO. It is perfectly straightforward until someone hands you one of those roof bricks that doesn’t have any of the attachment nodes on the top. Once you insert this brick, your tower can’t get any bigger.
If Xist did encode a protein, it would seem very odd that a cell would go to the effort of creating an RNA that was 17,000 bases[15] in length just to produce a protein that could have been encoded by an RNA of about 5 per cent of that length. Researchers in the field realised relatively quickly that this wasn’t what was happening. The reality was much stranger.
DNA is found in the nucleus. It’s copied to form RNA, and messenger RNA is transported out of the nucleus to structures where it acts as a template for protein assembly. But analyses showed that Xist RNA never left the nucleus. It doesn’t encode a protein, not even a short one.{121},{122}
Xist was in fact one of the first examples of an RNA molecule that is functional in its own terms, not as a carrier of information about a protein. It’s a great example of how junk DNA — DNA which doesn’t lead to production of a protein — is anything but junk. It’s extremely important in its own right, because without it X inactivation cannot happen.
An odd feature of Xist is not just that it doesn’t leave the nucleus. It doesn’t even leave the X chromosome that produces it. Instead, it essentially sticks to the inactive X and then spreads along the chromosome. As more and more Xist RNA is produced, it begins to spread out and cover the inactive X chromosome, in a process quaintly referred to as ‘painting’. The fact that this rather descriptive term is used is a quite good indicator that it’s something we don’t particularly understand. No one really knows the physical basis of how the Xist RNA creeps along the chromosome, like the mile-a-minute vine covering a wall. Even after more than twenty years we are still pretty hazy on how this happens. We do know that it’s not based on the sequence of the X chromosome. If the X inactivation centre is transferred on to an autosome in a cell, then the autosome can be inactivated as if it were an X.{123}
Although Xist is required to initiate the process of X inactivation, it has helpers that strengthen and maintain the process. As Xist paints the X chromosome, it acts as an attachment point for proteins in the nucleus. These bind to the inactivating X, and attract yet more proteins, which shut down expression even more tightly. The only gene that isn’t coated with Xist RNA and these proteins is the Xist gene itself. It remains a little beacon of expression in the chromosomal darkness of the inactive X.{124}
So we have here a situation where a piece of ‘junk’ DNA — one that doesn’t code for protein — is absolutely essential for the function of half the human race. Scientists have recently discovered that this process of X inactivation requires at least one other piece of junk DNA. Confusingly, this is encoded in exactly the same place on the X chromosome as Xist. DNA, as we know, is composed of two strands (the iconic double helix). The machinery that copies DNA to form RNA always ‘reads’ DNA in one direction, which we could call the beginning and end of a specific sequence. But the two strands of DNA run in opposite directions to each other, a little like one of those funicular railways we find at older seaside and mountain resorts. This means that a particular region of DNA may carry two lots of information in one physical location, running in opposite directions to each other.
A simple example in English is the word DEER, formed by reading from left to right. We could also read the same letters from right to left and in this case we would get the word REED. Same letters, different word, different meaning.
The other key piece of junk DNA involved in X inactivation is called, rather fittingly, Tsix. This is of course Xist spelt backwards, and it is found in the same region as Xist but on the opposite strand. Tsix encodes an RNA of 40,000 bases in length, over twice the size of Xist. Like Xist, Tsix never leaves the nucleus.
Although Tsix and Xist are encoded on the same part of the X chromosome, they are not expressed together. If an X chromosome expresses Tsix, this prevents the same chromosome from expressing Xist. This means that Tsix must be expressed by the active X chromosome, unlike Xist, which is always expressed from the inactive one.
This mutually exclusive expression of Tsix and Xist is of critical importance at a point in early development. The X chromosome in the egg has lost any of the protein marks that show it was inactivated (if it was the inactive version) and the X chromosome in the sperm had never been inactivated anyway. Following fusion and six or seven rounds of cell division, there will be a hundred or so cells in the embryo. At this stage, each cell in the female embryo switches off one of its two X chromosomes randomly. This requires a fleeting but intense physical relationship between the pair of X chromosomes in a cell. For just a couple of hours the two X chromosomes are physically associated in a brief encounter that ends with one being inactivated. The association is only over a small region of the X chromosome — the X inactivation centre, which codes for both Xist and Tsix RNA.{125}
This is the mother of all one-night stands. In those two hours, chromosomal decisions get made which are then maintained for the rest of life. Not just during foetal development, but right up until the woman dies, even if that is more than a hundred years later. And it affects not just the hundred or so cells, but the trillions that come after them, because the same X chromosome is inactivated in all daughter cells.
It’s still not entirely clear what happens during the hours of X chromosome intimacy in early development. The current theory is that there is a reallocation of junk RNA between the two chromosomes, such that one ends up with all the Xist and becomes the inactive X. We don’t know how, but it’s possible that one chromosome expresses slightly more or less of Xist or another key factor. We do know that the process begins just as levels of Tsix start to drop. It may be that once its levels fall below a certain critical threshold, Xist can start getting expressed from one of the X chromosomes.