The human X chromosome is a study in contradictions. The detailed sequence of the X, and a survey of inactivated genes in females, help to illuminate this unique ‘evolutionary space’.
The X chromosome was originally named X for ‘unknown’, remaining an oddity that has puzzled geneticists for centuries. It is the only chromosome to have one of the pair inactivated in one sex (females), and it is riddled with repeat elements, making it especially tough to produce a detailed gene sequence. From the papers by Ross et al.1 and Carrel and Willard2 (pages 325 and 400 of this issue), we learn that the wait was worth it.
The mammalian X and Y chromosomes arose from one pair of autosomes (that is, all the other chromosomes) 300 million years ago. Female mammals have two Xs; males have an X and a Y. After the X and Y seceded, mutations in genes on the Y made it the male-determining chromosome, and the pair began to diverge. The X and Y still share a few genes, mainly in the ‘pseudoautosomal’ region on the tip of the X; here the two chromosomes still exchange DNA to maintain proper segregation in cell division. Over time, the Y disintegrated to a shadow of its former self — but as long as the genes for maleness are preserved, most of the formerly autosomal genes seem to be largely extraneous because men already have one copy on the X.
In turn, the X developed a way to inactivate — silence — most of the genes on one of the two Xs in females, so that males and females would in large part have the same dosage of gene products. Early in female development, cells randomly choose either the maternal or paternal X to be the active X chromosome. The other one then transcribes large amounts of a large RNA from a gene in the middle of the chromosome, XIST. Only this chromosome becomes coated by XIST RNA, and thereby silenced by modification of its DNA and associated proteins. This choice is permanent, and has certain consequences. A famous example is the calico cat (Fig. 1). Similarly, human females are mosaics of the X chromosomes from each parent, and the severity of an X-linked disease in a female depends on the percentage of the cells in which the mutant gene concerned is silenced or expressed.
The inactive human X remains silenced even after being transferred into mouse cells, allowing for the formation of hybrid cells containing the inactive but not the active human chromosome. Carrel and Willard2 have exploited this system to create a near-complete catalogue of the inactivation state of X-linked genes. They find that 75% of genes are permanently silent, and about 15% permanently escape inactivation, meaning that they are expressed at twice the level in females as males. The remaining 10% are expressed in some inactive Xs but not others, indicating variability in human females that is likely to have medical relevance. Having twice the amount of any gene product could result in female–female or female–male differences: the possibilities, as yet undefined, are intriguing.
The mechanism of X-chromosome inactivation remains under debate, but the upshot is clear: in evolutionary terms, according to a principle known as Ohno's law, inactivation essentially froze the genes in place except in the pseudoautosomal regions. Yet comparative analyses of vertebrate genomic sequences tell us that human chromosomes have engaged in a big evolutionary swapping exercise, exchanging whole segments of DNA between them, even on the X and Y. The complete sequence of the gene-containing parts of the X and Y chromosomes1,3 has enabled Ross et al.1 to construct a remarkably detailed portrait of the X's past.
Previous comparisons between X- and Y-shared genes suggested that there are five evolutionary strata on the X, starting at the bottom of the long arm of the chromosome4. These strata represent large chromosome chunks formerly shared by X and Y — some original to the autosome pair, some arriving later — until inversions on one of the pair meant that X and Y could no longer exchange DNA in that region, moving the pseudoautosomal boundary upwards with each inversion. Selective pressure presumably maintains the small pseudoautosomal region that remains, so any further erosion of shared genes should be minimal.
X-chromosome inactivation is then proposed to work on each newly non-exchanging stratum, silencing genes to maintain proper dosage, while the Y version degrades over time. Indeed, Carrel and Willard2 found that the tendency to escape inactivation correlates with age of stratum, so that fewer genes escape in the oldest and more in the youngest. Despite all the shuffling, the X-inactivation centre remains at about the middle of the chromosome, also probably owing to selective pressure.
What genes have remained on the X? Because males have only one copy, diseases resulting from mutated X genes are often easier to identify in males — hence the early discoveries of genes involved in haemophilia and muscular dystrophy. According to previous dogma, the X is low in genes implicated in tumour formation and growth but high in those related to sex and reproduction5. It is surprising, then, that Ross et al.1 predict that a full 10% of protein-coding genes on the X produce ‘cancer-testis antigens’, a class of genes normally expressed mostly in the testis but whose activity is increased in testicular cancers, melanomas and other cancers.
These proteins are key targets for the development of cancer vaccines, as they are usually seen by the immune system only when they occur in tumours, and their expression is linked to the same mechanisms at work in X inactivation. As we know little about their normal function6, tailoring treatments will require a much better understanding of the role of the X in cancer. Given the prediction7 that genes on the X can benefit males even at the expense of females, as the detrimental effect would be masked in females, it may be that the genes encoding cancer-testis antigens convey some selective advantage to males, allowing some of the gene family to expand independently in both humans and mice1.
What about the sequences between the genes? The X chromosome is very rich in repeats (56% compared with a 45% average for the whole genome), and 14 gaps remain in the estimated 1.5-million base-pair sequence — despite the intensity of the sequencing effort1. But it's not just any repeats that matter. A whopping 29% of the chromosome consists of DNA repeats known as ‘long interspersed nuclear elements’ (LINEs), thought to be ‘selfish’ bits of DNA that jump around the genome. The main LINE constituents concerned are members of the L1 family and, in theory8, L1 repeats may serve as ‘booster elements’ for the X-inactivation signal, which emanates from around the middle of the chromosome with the transcription of XIST and spreads down the chromosome arms.
In support of this theory, the sequence shows almost no L1 elements in the XIST locus itself, but high concentrations on either side. L1 sequences are more common in the older areas of the chromosome, with more genes subject to inactivation, and less frequent in the newer areas, where more genes escape1,2. So it is possible that more genes may become subject to inactivation over time, as still-active L1 elements invade in higher numbers.
The exact mechanism for L1 involvement is not known: is this ‘junk DNA’ crucial in shaping the X, or merely the footprint left behind as the process progresses? Mice and cows do not show similar patterns of LINE elements on their Xs, or of any other known repeat element that might fit the bill, yet they have many similar features of X-chromosome inactivation. Factors involved in inactivating the X are modifications to the DNA and proteins in the chromosome, and formation of a complex with XIST RNA and presumably associated proteins. Intensive research is aimed at finding out how these modifications might have evolved, and how they now establish and maintain the inactivated state.
Like the rest of the human genome, the primary sequence of this chromosome is an advanced starting point for exploring the mysteries of evolution and development. Geneticists eagerly await genomic sequences from non-placental mammals, such as the platypus and opossum, so as to reconstruct earlier, proto-X chromosomes. As the rock band U2 would say, the X moves in mysterious ways, and we've just been given a preview.
Ross, M. T. et al. Nature 434, 325–337 (2005).
Carrel, L. & Willard, H. F. Nature 434, 400–404 (2005).
Skalestky, H. et al. Nature 423, 825–837 (2003).
Lahn, B. T. & Page, D. C. Science 286, 964–967 (1999).
Kohn, M. et al. Trends Genet. 20, 598–603 (2004).
Spatz, A. et al. Nature Rev. Cancer 4, 617–629 (2004).
Rice, W. R. Evolution 38, 735–742 (1984).
Lyon, M. F. Cytogenet. Cell Genet. 80, 133–137 (1998).